Methods, compositions, and kits suitable for inflammatory bowel disease (IBD) and IBD subtypes

Abstract

Methods are provided for determining, diagnosing, preventing, or treating Crohn's disease (CD), ulcerative colitis (UC), and / or inflammatory bowel disease (IBD) in an individual. Some embodiments provide kits and computer program products for determining or diagnosing Crohn's disease (CD), ulcerative colitis (UC), and / or inflammatory bowel disease (IBD) in an individual. Other exemplary embodiments are described herein. In some embodiments, the disclosed methods and kits are accurate, cost-effective, and non-invasive.

Description

Cross-references to related applications

[0001] This application claims priority and interest in U.S. Provisional Application No. 63 / 562,232, filed March 6, 2024; U.S. Provisional Application No. 63 / 562,233, filed March 6, 2024; U.S. Provisional Application No. 63 / 675,266, filed July 25, 2024; and U.S. Provisional Application No. 63 / 689,864, filed September 3, 2024. All contents of the foregoing applications are hereby incorporated herein by reference in their entirety for all purposes. Reference to the sequence list

[0002] This application contains a sequence list that has been filed electronically in ST.26 (xml) format and is hereby incorporated by reference in its entirety. The copy of said ST.26, created on March 4, 2025, is named “M057001PCTCN.xml” and is 182 kilobytes in size. Technical Field

[0003] This application relates to methods and kits for identifying, diagnosing, preventing, or treating Crohn's disease (CD), ulcerative colitis (UC), and / or inflammatory bowel disease (IBD) in an individual; and to compositions for maintaining health, preventing, or treating Crohn's disease (CD), ulcerative colitis (UC), and / or inflammatory bowel disease (IBD) in an individual. Background Technology

[0004] Inflammatory bowel disease (IBD), including Crohn's disease (CD) and ulcerative colitis (UC), is a chronic and relapsing inflammatory disorder of the gastrointestinal tract (GI). It is estimated that more than seven million people worldwide suffer from IBD. Although IBD was historically more prevalent in Western countries, newly industrialized nations have experienced a rise in disease incidence over the past few decades, primarily attributed to lifestyle and environmental factors. Delayed diagnosis is often associated with disease progression and bowel surgery, while early diagnosis and timely intervention can improve prognosis.

[0005] Ulcerative colitis (UC) is a chronic inflammatory bowel disease that primarily affects the rectum and colon. Crohn's disease (CD) is another subtype of inflammatory bowel disease that can lead to progressive bowel damage and disability. The exact cause of CD remains unknown. Currently, there is no single reference standard for the diagnosis of CD. Diagnosis of both CD and UC typically requires a combination of invasive procedures, such as colonoscopy, esophagogastric-duodenoscopy (OGD), and computed tomography (CT). Therefore, new / alternative methods are needed to determine the risk, diagnose, prevent, or treat CD, UC, and / or IBD. Summary of the Invention

[0006] In light of the foregoing background, in some embodiments, the objective is to provide novel kits, methods, and uses for determining, diagnosing, preventing, or treating Crohn's disease (CD), ulcerative colitis (UC), and / or inflammatory bowel disease (IBD).

[0007] In some embodiments, a method is provided for determining, diagnosing, preventing, or treating inflammatory bowel disease (IBD) in an individual, the method comprising the step of: a) determining the relative abundance of a set of bacterial species in a fecal sample from the individual, wherein the set of bacterial species comprises one or more bacterial species selected from the group consisting of: Actinomyces sp. oral taxon 181, Bacteroides fragilis, Escherichia coli, Lawsonibacter asaccharolyticus, Eubacterium sp. CAG: 274, Roseburia inulinivorans, Roseburia intestinalis, Ruminococcus obeum (Blautia obeum), and Dorea... a) The relative abundance of each bacterial species in the bacterial species set from the individual is compared with the relative abundance of each bacterial species in the bacterial species set in the reference dataset to generate a risk score; c) If the risk score is higher than a cutoff value, the individual is determined to have an increased risk of IBD or currently has IBD; and d) If it is determined that the individual has an increased risk of developing IBD or already has IBD, the individual may optionally be treated; wherein the IBD is active or inactive IBD.

[0008] In some embodiments, a method for determining, diagnosing, preventing, or treating inflammatory bowel disease (IBD) and / or IBD subtypes, wherein the IBD subtype is ulcerative colitis (UC) or Crohn's disease (CD), the method comprising the steps of: (a) determining the relative abundance of a first set of bacterial species, optionally fecal calprotectin levels, and optionally the relative abundance of a second set of bacterial species in a fecal sample from the individual, wherein the first set of bacterial species and the second set of bacterial species each independently comprise one or more bacterial species selected from the group consisting of: Actinomycetes species oral taxa 181, Bacteroides fragilis, Escherichia coli, Lawsonia suisin, and Eubacterium species CAG. 274. *Roseobacterium inulinum*, *Roseobacterium ovale*, *Broutella ovale*, *Dolphobicella formate*, *Biliobacterium warfarinii*, *Clostridium tumefaciens*, *Clostridium sacchariformis*, *Gemmelella formica*, *Ostomyces viscerans*, *Ruminococcus tortifolia*, *Clostridium spiralis*, *Twincocephalus measles*, and *Broutella hensenii*; (b) comparing the relative abundance of each bacterial species from the first bacterial species set of the individual with the relative abundance of each bacterial species from the first bacterial species set in the first reference dataset, and optionally comparing the fecal calprotectin level from the individual with the fecal calprotectin level in the first reference dataset using a first machine learning model to generate a first risk score; (c) if the first risk score is higher than a first cutoff value, determining that the individual has an increased risk of IBD or currently has IBD; (d) If it is determined that the individual has an increased risk of developing IBD or has IBD, then optionally a second machine learning model is used to compare the relative abundance of each bacterial species in the second bacterial species set from the individual with the relative abundance of each bacterial species in the second bacterial species set in the second reference dataset to generate a second risk score; (e) optionally, if the second risk score is higher than a second cutoff value, then it is determined that the individual has an increased risk of developing CD or has CD, or if the risk score is lower than or equal to the second cutoff value, then it is determined that the individual has an increased risk of developing UC or has UC; and (f) if it is determined in step (c) that the individual has an increased risk of developing IBD or has IBD, or if steps (d) and (e) are performed and it is determined that the individual has an increased risk of developing UC or CD or has UC or CD, then optionally the individual is treated.

[0009] In some embodiments, a method is provided for diagnosing, preventing, or treating an IBD subtype in an individual, wherein the IBD subtype is ulcerative colitis (UC) or Crohn's disease (CD), the method comprising the steps of: a) determining whether the individual has an increased risk of developing IBD or currently has IBD; b) if the individual is determined to have an increased risk of developing IBD or currently has IBD, determining the relative abundance of a second set of bacterial species in a fecal sample from the individual, wherein the second set of bacterial species comprises one or more bacterial species selected from the group consisting of: Actinomycetes species oral taxa 181, Bacteroides fragilis, Escherichia coli, Lawsonia suisinophila, and Eubacterium species CAG. 274. *Roseobacterium inulinum*, *Roseobacterium brevesii*, *Ruminococcus ovalis* (*Broutella ovalis*), *Dolphobicella formate*, *Biliobacterium warfarinii*, *Clostridium tumefaciens*, *Clostridium sacchariformis*, *Gemmelella formica*, *Ostomyces viscerans*, *Ruminococcus tortifolia*, *Clostridium spiralis*, *Twincocephalus measles*, and *Broutella hensenii*; c) using a second machine learning model to compare the relative abundance of each bacterial species in the second bacterial species set from the individual with the relative abundance of each bacterial species in the second bacterial species set in the second reference dataset to generate a second risk score; d) if the second risk score is higher than a second cutoff value, the individual is determined to have an increased risk of having CD or currently has CD, or if the risk score is lower than or equal to the second cutoff value, the individual is determined to have an increased risk of having UC or currently has UC; and e) if the individual is determined to have an increased risk of having UC or CD or currently has UC or CD, the individual is optionally treated.

[0010] In some embodiments, a method is provided for identifying a fecal sample with an altered inflammatory bowel disease (IBD) microbiome and / or an altered IBD subtype microbiome, wherein the IBD subtype is ulcerative colitis (UC) or Crohn's disease (CD), the method comprising the steps of: a) determining the relative abundance of a first set of bacterial species, optionally fecal calprotectin levels, and optionally the relative abundance of a second set of bacterial species in the fecal sample, wherein the first set of bacterial species and the second set of bacterial species each independently comprise one or more bacterial species selected from the group consisting of: Actinomycetes species oral taxa 181, Bacteroides fragilis, Escherichia coli, Lawsonia suisin, and Eubacterium species CAG. 274. *Roseidobacterium ovale*, *Roseidobacterium enterica*, *Broutella ovale*, *Dolphobicella formate*, *Biliobacterium wartii*, *Clostridium sacchariformis*, *Clostridium formicum*, *Clostridium viscerum*, *Ruminococcus tortifolium*, *Clostridium spiralis*, *Twincocephalus measles*, and *Broutella hensenii*; b) The relative abundance of each bacterial species in the first bacterial species set from the fecal sample is compared with the relative abundance of each bacterial species in the first bacterial species set in the first reference dataset, and optionally, a first machine learning model is used to compare the fecal calprotectin level from the fecal sample with the fecal calprotectin level in the first reference dataset to generate a risk score; and c) if the first risk score is higher than a first cutoff value, the fecal sample is determined to have an altered IBD microbiome; d) if the fecal sample is determined to have an altered IBD microbiome, optionally, a second machine learning model is used to compare the relative abundance of each bacterial species in the second bacterial species set from the fecal sample with the relative abundance of each bacterial species in the second bacterial species set in the second reference dataset to generate a second risk score; and e) optionally, if the second risk score is higher than a second cutoff value, the fecal sample is determined to have an altered CD microbiome, or if the risk score is lower than or equal to the second cutoff value, the fecal sample is determined to have an altered UC microbiome.

[0011] In some embodiments, a method is provided for identifying a fecal sample with an altered IBD subtype microbiome, wherein the IBD subtype is ulcerative colitis (UC) or Crohn's disease (CD), the method comprising the steps of: a) determining whether the fecal sample has an altered IBD microbiome; b) if the fecal sample is determined to have an altered IBD microbiome, determining the relative abundance of a second set of bacterial species in the fecal sample from the individual, wherein the second set of bacterial species comprises one or more bacterial species selected from the group consisting of: Actinomycetes species oral taxa 181, Bacteroides fragilis, Escherichia coli, Lawsonia sucrase, Eubacterium species CAG: 274, Rosette spp., Rosette spp., Oval rumenococci (Oval Brutella), Formic acid-producing Dorl, Bileophilus warwickii, Clostridium spp., Clostridium spp., Formic acid-producing Clostridium, Formic acid-producing Gemmaie, Visceral Odorobacter, Ruminococcus tortiformis, Clostridium spiralis, Gemmaietta spp., and Broutella henselae; c) The relative abundance of each bacterial species in the second bacterial species set from the fecal sample is compared with the relative abundance of each bacterial species in the second bacterial species set in the second reference dataset using a second machine learning model to generate a second risk score; and d) if the second risk score is higher than a second cutoff value, the fecal sample is determined to have an altered CD microbiome, or if the risk score is lower than or equal to the second cutoff value, the fecal sample is determined to have an altered UC microbiome.

[0012] In some embodiments, a kit is provided for determining the risk or diagnosing inflammatory bowel disease (IBD) and / or IBD subtypes in an individual, wherein the IBD subtype is ulcerative colitis (UC) or Crohn's disease (CD). The kit comprises reagents for detecting a first set of bacterial species and optionally reagents for detecting a second set of bacterial species, wherein the first set of bacterial species and the second set of bacterial species each independently comprise one or more bacterial species selected from the group consisting of: Actinomycetes species oral taxa 181, Bacteroides fragilis, Escherichia coli, Lawsonia sucrase, Eubacterium species CAG: 274, Rosette of inulin, Rosette of the gut, Oval rumenococcus (Oval Brutella), Formic acid-producing Dorl, Bileophilus warwickii, Clostridium truncatella, Clostridium sucrase, Formic acid-producing Gemmaella, Visceral Odorobacter, Ruminococcus tortiformis, Clostridium spiralis, Gemmaella muscarinae, and Broutella hensenae.

[0013] In some embodiments, a computer program product is provided for determining the risk of or diagnosing inflammatory bowel disease (IBD) and / or IBD subtypes in an individual, wherein the IBD subtype is ulcerative colitis (UC) or Crohn's disease (CD), wherein the computer program product comprises a computer-readable medium encoded with computer-executable code configured to perform the steps of the method as described in any one of Examples 2-47.

[0014] In some embodiments, a computer program product is provided for determining, diagnosing, or recommending treatment for inflammatory bowel disease (IBD) and / or IBD subtypes, wherein the IBD subtype is ulcerative colitis (UC) or Crohn's disease (CD), wherein the computer program product comprises a computer-readable medium encoded with computer-executable code configured to perform the steps of the method as described in any one of Examples 2-47, and optionally further configured to perform the following steps: recommending appropriate IBD treatment, UC treatment, or CD treatment to the individual if it is determined that the individual has an increased risk of developing IBD, UC, or CD, or is currently suffering from IBD, UC, or CD.

[0015] In some embodiments, a computer program product is provided for determining the risk or diagnosing an IBD subtype in an individual, wherein the IBD subtype is ulcerative colitis (UC) or Crohn's disease (CD), wherein the computer program product comprises a computer-readable medium encoded with computer-executable code configured to perform the steps of the method as described in any one of Examples 48-78.

[0016] In some embodiments, a computer program product is provided for determining, diagnosing, or recommending treatment for an individual for an IBD subtype, wherein the IBD subtype is ulcerative colitis (UC) or Crohn's disease (CD), wherein the computer program product comprises a computer-readable medium encoded with computer-executable code, wherein the computer-executable code is configured to perform the steps of the method as described in any one of Examples 48-78, and optionally further configured to perform the following steps: recommending appropriate IBD treatment, UC treatment, or CD treatment to the individual, respectively, if it is determined that the individual has an increased risk of developing IBD, UC, or CD, or is currently suffering from IBD, UC, or CD.

[0017] In some embodiments, a method is provided for determining, diagnosing, preventing, or treating Crohn's disease (CD) in an individual, the method comprising the steps of: a) determining the relative abundance of a set of bacterial species in a fecal sample from the individual, wherein the set of bacterial species comprises one or more bacterial species selected from the group consisting of: *Dolphus formateriate*, *Eubacteria eligens*, *Roseobacterium inulinum*, *Roseobacterium enterica*, *Ruminococcus ovalis* (*Broutella ovalis*), *Eubacterium ventriosum*, *Anaerostipes hadrus*, *Lawsonia suisophilus*, *Oscillibacter* species CAG: 241, *Oscillibacter* species 57-20, *Lactobacillus rogosae*, *Eubacterium* species CAG: 274, *Romboutsia*. (b) The relative abundance of each bacterial species in the bacterial species set from the individual is compared with the relative abundance of each bacterial species in the bacterial species set in the reference dataset to generate a risk score; (c) If the risk score is higher than a cutoff value, the individual is determined to have an increased risk of developing CD or to have CD; and (d) If the individual is determined to have an increased risk of developing CD or to have CD, the individual is optionally treated.

[0018] In some embodiments, a method is provided for identifying a fecal sample with an altered Crohn's disease microbiome, the method comprising the steps of: a) determining the relative abundance of a set of bacterial species in the fecal sample, wherein the set of bacterial species comprises one or more bacterial species selected from the group consisting of: *Dolphobic formate*, *Eubacterium pickyii*, *Roseinia inulinica*, *Roseinia enterica*, *Ruminococcus ovalis* (*Broutella ovalis*), *Eubacterium truncatella*, *Anaerobic bacillus hardruciformis*, *Lawsonia solani*, *Oscillatoria* species CAG: 241, *Oscillatoria* species 57-20, *Lactobacillus rhamnosus*, *Eubacterium* species CAG: 274, *Romebuts ileum*, *Bacteroides fragilis*, *Escherichia coli*, *Eubacterium gingivalis*, *Actinomyces* species oral taxonomy group 181, and *Actinomyces* species S6-Spd3; b) The relative abundance of each bacterial species in the bacterial species set from the individual is compared with the relative abundance of each bacterial species in the bacterial species set in the reference dataset to generate a risk score; and c) if the risk score is higher than a cutoff value, the fecal sample is determined to have an altered CD microbiome.

[0019] In some embodiments, a method is provided to increase amino acid biosynthesis in cells, wherein the amino acid is L-arginine, L-ornithine, or L-valine, the method comprising providing the cells with an effective amount of one or more bacterial species selected from the group consisting of: ovale-forming rumenococci (ovale-forming broutitidis), inulin-eating Rosette, formic acid-producing Dorperella, Eubacterium species CAG: 274, and Enterobacter Rosette.

[0020] In some embodiments, a method for increasing carbohydrate degradation in cells is provided, the method comprising providing the cells with an effective amount of one or more bacterial species selected from the group consisting of: ovale-forming rumenococci (ovale-forming Brontë), inulin-eating Rosette, formic acid-producing Dorper, Eubacterium species CAG: 274, and Enterobacter Rosette.

[0021] In some embodiments, a method is provided for increasing the biosynthesis of cofactors, carriers, and vitamins in cells, wherein the vitamin is thiamine pyrophosphate, the method comprising providing the cells with an effective amount of one or more bacterial species selected from the group consisting of: ovale-forming rumenococci (ovale-forming Broutella), inulin-eating Rosette, formic acid-producing Dorperella, Eubacterium species CAG: 274, and Enterobacter Rosette.

[0022] In some embodiments, a method is provided to increase amino acid biosynthesis in cells, wherein the amino acid is L-tryptophan, the method comprising providing the cells with an effective amount of ovoid rumenococci (ovoid broutella), inulin-eating Rosette, formic acid-producing Dorperella, Eubacterium species CAG: 274, or Enterobacter Rosette.

[0023] In some embodiments, a composition is provided comprising Lawsonia non-saccharitroph and Eubacterium species CAG: 274.

[0024] In some embodiments, a composition is provided comprising ovoid rumenococci (ovoid Broughtus), sugar-insoluble Lawsonia, and inulin-eating Rosette.

[0025] In some embodiments, a kit is provided for determining the risk of or diagnosing Crohn's disease (CD) in an individual, the kit comprising reagents for detecting a set of bacterial species selected from the group consisting of: *Dolphus formateriate*, *Eubacterium pickyii*, *Roseobacterium inulinum*, *Roseobacterium enterica*, *Ruminococcus ovalis*, *Eubacterium truncatula*, *Anaerobic bacillus hardruciformis*, *Lawsonia suis*, *Oscillatoria* species CAG: 241, *Oscillatoria* species 57-20, *Lactobacillus rhamnosus*, *Eubacterium* species CAG: 274, *Romebuts ileum*, *Bacteroides fragilis*, *Escherichia coli*, *Eubacterium gingivalis*, *Actinomyces* species oral taxa 181, and *Actinomyces* species S6-Spd3.

[0026] In some embodiments, a computer program product is provided for determining the risk of or diagnosing Crohn's disease (CD) in an individual, wherein the computer program product comprises a computer-readable medium encoded with computer-executable code, wherein the computer-executable code is configured to perform the following steps: a) determining the relative abundance of a set of bacterial species in a fecal sample from the individual, wherein the set of bacterial species comprises one or more bacterial species selected from the group consisting of: *Dolphobic formate*, *Eubacterium pickyii*, *Roseobacterium inulinum*, *Roseobacter rumenella*, *Broutella ovalis*, *Eubacterium truncatella*, *Anaerobic bacillus hardruciformis*, *Lawsonia suis*, *CAG: 241*, *CAG: 57-20*, *Lactobacillus rhamnosus*, *CAG: 274*, *Romebutzella ileum*, *Bacteroides fragilis*, *Escherichia coli*, *Eubacterium gingivalis*, *Actinomyces* species oral taxonomy group 181, and *Actinomyces* species S6-Spd3; b) The relative abundance of each bacterial species in the bacterial species set from the individual is compared with the relative abundance of each bacterial species in the bacterial species set in the reference dataset to generate a risk score; and c) if the risk score is higher than a cutoff value, the individual is determined to have an increased risk of CD or to have CD.

[0027] In some embodiments, a method is provided for determining, diagnosing, preventing, or treating Crohn's disease (CD) in an individual, the method comprising the steps of: a) determining the relative abundance of a set of bacterial species in a fecal sample from the individual, wherein the set of bacterial species comprises one or more bacterial species selected from the group consisting of: *Gemisia formate*, *Eubacterium hallii*, *Broutella ovalis*, *Roseobacterium inulinum*, *Clostridium sacchariformis*, *Eubacterium rectale*, *Roseburia faecalis*, *Asaccharobacter celatus*, *Collinsella aerofaciens*, *Faecalibacterium prausnitzii*, *Anaerobic Proteobacterium hardrovirus*, *Lachnospira pectinoschiza*, *Ruminococcus tortifolia*, *Clostridium truncatella*, and *Parabacteroides*. *Merdae*, *Ruminococcus bromii*, *Rossella enterica*, *Adlercreutzia equolifaciens*, *Alistipes putredinis*, *Eubacterium* species CAG: 38, *Roseburia hominis*, *Agathobaculum butyriciproductns*, *Dorea formicate-producing *Bacteroides stercoris*, *Visceral Osmophilus*, *Oscillatoria* species 57_20, *Coprococcus catus*, *Bacteroides vulgatus*, *Dorea longicatena*, *Alistipes shahii*, *Lawsonia unsacchariformis*, *Firmicutes bacterium* CAG: 83, *Oscillatoria* species CAG: 241. *Eubacterium ramulus*, *Alistipes finegoldii*, *Bacteroides caccae*, *Akkermansia muciniphila*, *Coprococcus comes*, *Butyricimonas virosa*, *Alistipes**Indistinctus*, *Bacteroides uniformis*, *Bacteroides coprocola*, *Eubacterium* species CAG: 274, *Desulfovibrio piger*, *Bacteroides massiliensis*, *Clostridium sp.* species CAG: 299, *Piperella spp.*, *Romebuts ileum*, *Eubacterium spp.*, *Anaerobic bacillus hardruciformis*, *Lactobacillus rogues*, *Ruminococcus ovalis*, *Gemella haemolysans*, *Streptococcus mitis*, *Actinomyces odontolyticus*, *Actinomyces* species S6-Spd3, *Actinomyces graevenitzii*, *Lactobacillus mucosa*, *Streptococcus parasanguinis*, *Rothia* a) The relative abundance of each bacterial species in the bacterial species set from the individual is compared with the relative abundance of each bacterial species in the bacterial species set in the reference dataset to generate a risk score; c) If the risk score is higher than a cutoff value, the individual is determined to have an increased risk of CD or to have CD; and d) If the individual is determined to have an increased risk of CD or to have CD, the individual is optionally treated.

[0028] In some embodiments, a method is provided for monitoring Crohn's disease (CD) activity in an individual, the method comprising the steps of: a) determining the relative abundance of a set of bacterial species in a fecal sample from the individual, wherein the set of bacterial species comprises one or more bacterial species selected from the group consisting of: formic acid-producing Dorperella, ovale-forming rumenococci (ovale-forming Broutella), sugar-insoluble Lawsonia, and Escherichia coli; b) comparing the relative abundance of each bacterial species in the set of bacterial species from the individual with the relative abundance of each bacterial species in the set of bacterial species in a reference dataset to generate a risk score; c) if the risk score is higher than a cutoff value, determining that the individual has increased disease activity.

[0029] In some embodiments, a method is provided for determining, diagnosing, preventing, or treating ulcerative colitis (UC) in an individual, the method comprising the steps of: a) determining the relative abundance of a set of bacterial species in a fecal sample from the individual, wherein the set of bacterial species comprises one or more bacterial species selected from the group consisting of: *Phascolarctobacterium faecium*, *Phascolarctobacterium crypticum*, *Collinsella stercoris*, *Vibrio vulnificus* species CAG: 241, *Lawsonia sucraniosus*, *Butymonas vulnificus*, *Clostridium* species CAG: 58, *Eubacterium* species CAG: 274, *Biliobacterium warfarinii*, *Clostridium truncatum*, *Clostridium sacchariformis*, *Gemmelia formica*, *Ostomyces viscerans*, *Ruminococcus tortifolius*, *Actinomyces* species oral taxonomy group 181, *Clostridium spiralis*, *Twinococcus mesenteriae*, and *Broutella henselae*; b) The relative abundance of each bacterial species in the bacterial species set from the individual is compared with the relative abundance of each bacterial species in the bacterial species set in the reference dataset to generate a risk score; c) if the risk score is higher than a cutoff value, the individual is determined to have an increased risk of UC or to have UC; and d) if the individual is determined to have an increased risk of UC or to have UC, the individual is optionally treated.

[0030] In some embodiments, a method is provided for identifying a fecal sample with an altered ulcerative colitis microbiome, the method comprising the steps of: a) determining the relative abundance of a set of bacterial species in the fecal sample, wherein the set of bacterial species comprises one or more bacterial species selected from the group consisting of: *Phascolarctobacterium faecium*, *Phascolarctobacterium cryptana*, *Collinsella stercoris*, *Vibrio vulnificus* species CAG: 241, *Lawsonia cryptana*, *Butymonas vulnificus*, *Clostridium* species CAG: 58, *Eubacterium* species CAG: 274, *Bacillus warfarinii*, *Clostridium truncatula*, *Clostridium sacchariformis*, *Gemmelia formica*, *Ostomyces viscerans*, *Ruminococcus tortifolius*, *Actinomyces* species oral taxonomy group 181, *Clostridium spiralis*, *Twinococcus mesenteriae*, and *Broutella henselae*; b) The relative abundance of each bacterial species in the bacterial species set from the individual is compared with the relative abundance of each bacterial species in the bacterial species set in the reference dataset to generate a risk score; and c) if the risk score is higher than a cutoff value, the fecal sample is determined to have an altered UC microbiome.

[0031] In some embodiments, a method is provided to increase nucleoside and nucleotide biosynthesis in cells, the method comprising providing the cells with an effective amount of one or more bacterial species selected from the group consisting of: Clostridium truncatum, Clostridium sacchariformis, Gemmaella formica, and Ruminococcus tortifolia.

[0032] In some embodiments, a composition is provided comprising Clostridium sacchariformis, Clostridium truncatum, and Formic acid-producing Gastroenterobacter.

[0033] In some embodiments, a kit is provided for determining the risk of or diagnosing ulcerative colitis (UC) in an individual, the kit comprising reagents for detecting a set of bacterial species selected from the group consisting of: *Collagenia fecalis*, *Bacillus crypticarum*, *Colinus feces*, *Oscillatoria* species CAG:241, *Lawsonia crypticarum*, *Toxococcus butyricum*, *Clostridium* species CAG:58, *Eubacterium* species CAG:274, *Bacillus warwickii*, *Clostridium tenuifolium*, *Clostridium sacchariformis*, *Gemmelia formica*, *Ostomyces viscerans*, *Ruminococcus tortifolia*, *Actinomyces* species oral taxonomy group 181, *Clostridium spiralis*, *Twinia measles*, and *Brutella henselae*.

[0034] In some embodiments, a computer program product is provided for determining the risk of or diagnosing ulcerative colitis (UC) in an individual, wherein the computer program product comprises a computer-readable medium encoded with computer-executable code, wherein the computer-executable code is configured to perform the following steps: a) determining the relative abundance of a set of bacterial species in a fecal sample from the individual, wherein the set of bacterial species comprises one or more bacterial species selected from the group consisting of: *Coalabacterium feces*, *Bacillus spp.*, *Clinobacterium feces*, *Oscillatoria* species CAG: 241, *Lawsonia spp.*, *Toxococcus butyricum*, *Clostridium* species CAG: 58, *Eubacterium* species CAG: 274, *Bacillus warfarinii*, *Clostridium spp.*, *Clostridium spp.*, *Clostridium spp.*, *Gemmelia formica*, *Ostomyces viscerans*, *Ruminococcus tortifolius*, *Actinomyces* species oral taxonomy group 181, *Clostridium spiralis*, *Twinococcus mesenteriae*, and *Broutella henselae*; b) The relative abundance of each bacterial species in the bacterial species set from the individual is compared with the relative abundance of each bacterial species in the bacterial species set in the reference dataset to generate a risk score; and c) if the risk score is higher than a cutoff value, the individual is determined to have an increased risk of UC or to have UC.

[0035] In some embodiments, a method is provided for determining, diagnosing, preventing, or treating ulcerative colitis (UC) in an individual, the method comprising the steps of: a) determining the relative abundance of a set of bacterial species in a fecal sample from the individual, wherein the set of bacterial species comprises one or more bacterial species selected from the group consisting of: *Colinus oxytetracyclis*, *Clostridium truncatella*, *Ruminococcus tortifolius*, *Bacillus sacchariformis*, *Gemacobacterium formica*, *Clostridium sacchariformis*, *Alternaria putrefactiveis*, *Dolphinium longiformis*, *Ruminococcus brevicornu*, *Ostomyces viscerans*, *Trichophyton floccosum*, *Concomitant cocci*, *Adlerococcus adrenocorticotropicus*, *Roseidonella inulinophila*, *Brutella ovalis*, *Eubacterium rectum*, *Dolphinium formica*, *Alternaria sarcopenia*, *Bacteroides fecalithana*, *Bacteroides fecalithana*, and *CAG* species. 241. *Ackermania pseudomallei*, *Alternaria fibrillosa*, *Eubacterium hominis*, *Oscillatoria* species 57_20, *Rossiella humanis*, *Koala faecalis*, *Collinus faecalis*, *Bifidobacterium adolescentis*, *Rossiella enterica*, *Toxic Butymonas*, *Alternaria unobservatus*, *Eubacterium* species CAG: 38, *Anaerobic Bacteroides harlequinae*, *Bacteroides faecalis*, *Clostridium* species CAG: 242, *Lawsonia suisinensis*, *Eubacterium* species CAG: 274, *Cercozoella fecalithii*, *Bacteroides masei*, *Rossiella faecalis*, *Clostridium* species CAG: 58, *Ruthenibacterium lactatiformans*, *Bacteroides plebeius*, *Biliophilus warwickii*, Firmicutes bacteria CAG: 145, *Eubacterium clematis*, *Enterorhabdus cecum* a) The relative abundance of each bacterial species in the bacterial species set from the individual is compared with the relative abundance of each bacterial species in the bacterial species set in the reference dataset to generate a risk score; c) If the risk score is higher than a cutoff value, the individual is determined to have an increased risk of UC or currently has UC; and d) If the individual is determined to have an increased risk of developing UC or is already suffering from UC, the individual may optionally be treated.

[0036] In some embodiments, a method is provided for monitoring ulcerative colitis (UC) disease activity in an individual, the method comprising the steps of: a) determining the relative abundance of a set of bacterial species in a fecal sample from the individual, wherein the set of bacterial species comprises one or more bacterial species selected from the group consisting of: Clostridium truncatum, Clostridium sacchariformis, Osmotherium visceratum, Gemmaella formica, Actinomyces species oral taxonomy 181, and Clostridium spiralis; b) comparing the relative abundance of each bacterial species in the set of bacterial species from the individual with the relative abundance of each bacterial species in the set of bacterial species in a reference dataset to generate a risk score; and c) determining that the individual has increased disease activity if the risk score is higher than a cutoff value.

[0037] In some embodiments, a computer program product is provided for determining the risk, diagnosing, or recommending treatment for Crohn's disease (CD) in an individual, wherein the computer program product comprises a computer-readable medium encoded with computer-executable code, wherein the computer-executable code is configured to perform the following steps: a) determining the relative abundance of a set of bacterial species in a fecal sample from the individual, wherein the set of bacterial species comprises one or more bacterial species selected from the group consisting of: *Dolphobic formate*, *Eubacterium pickyii*, *Roseobacterium inulinum*, *Roseobacterium enterica*, *Ruminococcus ovalis* (*Broutella ovalis*), *Eubacterium truncatula*, *Anaerobic bacillus hardruciformis*, *Lawsonia suis*, *Oscillatoria* species CAG: 241, *Oscillatoria* species 57-20, *Lactobacillus rhamnosus*, *Eubacterium* species CAG: 274, *Romebuts ileum*, *Bacteroides fragilis*, *Escherichia coli*, *Eubacterium gingivalis*, *Actinomyces* species oral taxonomy group 181, and *Actinomyces* species S6-Spd3; b) The relative abundance of each bacterial species in the bacterial species set from the individual is compared with the relative abundance of each bacterial species in the bacterial species set in the reference dataset to generate a risk score; c) if the risk score is higher than a cutoff value, the individual is determined to have an increased risk of developing CD or to have CD; and d) if the individual is determined to have an increased risk of developing CD or to have CD, appropriate CD treatment is optionally recommended to the individual.

[0038] In some embodiments, a computer program product is provided for determining the risk of or diagnosing ulcerative colitis (UC) in an individual, wherein the computer program product comprises a computer-readable medium encoded with computer-executable code, wherein the computer-executable code is configured to perform the following steps: a) determining the relative abundance of a set of bacterial species in a fecal sample from the individual, wherein the set of bacterial species comprises one or more bacterial species selected from the group consisting of: *Coalabacterium feces*, *Bacillus spp.*, *Clinobacterium feces*, *Oscillatoria* species CAG: 241, *Lawsonia spp.*, *Toxococcus butyricum*, *Clostridium* species CAG: 58, *Eubacterium* species CAG: 274, *Bacillus warfarinii*, *Clostridium spp.*, *Clostridium spp.*, *Clostridium spp.*, *Gemmelia formica*, *Ostomyces viscerans*, *Ruminococcus tortifolius*, *Actinomyces* species oral taxonomy group 181, *Clostridium spiralis*, *Twinococcus mesenteriae*, and *Broutella henselae*; b) The relative abundance of each bacterial species in the bacterial species set from the individual is compared with the relative abundance of each bacterial species in the bacterial species set in the reference dataset to generate a risk score; c) if the risk score is higher than a cutoff value, the individual is determined to have an increased risk of UC or to have UC; and if the individual is determined to have an increased risk of UC or to have UC, appropriate UC treatment is optionally recommended to the individual.

[0039] In some embodiments, a computer-readable medium is provided, the medium including instructions that, when executed by at least one computer, cause the at least one computer to perform the steps of a method according to any one of the examples described herein.

[0040] In some embodiments, a computer system is provided, the system including at least one memory, at least one processor, and a computer program stored in the at least one memory, wherein the at least one processor executes the computer program to perform steps of a method according to any one of the examples described herein.

[0041] Other exemplary embodiments will be described below. advantage

[0042] This disclosure has many advantages.

[0043] In some embodiments, this disclosure provides novel methods and kits for assessing risk and diagnosing IBD in subjects, and / or identifying / diagnosing an IBD subtype (CD or UC) in said subjects using one or more bacterial biomarkers. By determining the relative abundance of these bacterial biomarkers in fecal samples and generating risk or diagnostic scores, the disclosed techniques can provide a cost-effective, non-invasive approach to support clinical decision-making, thereby contributing to improved disease management. For example, the choice of medical treatment varies depending on the disease subtype. UC and CD have different pathogenesis and differentiated responses to drug treatment. Medical treatments often differ depending on disease severity and / or whether a progressive approach is used. For example, in mild to moderate UC, 5-aminosalicylic acid (5-ASA) is the recommended treatment. Biologics are commonly used to treat moderate to severe CD and UC, but some drugs are more targeted at one disease. For example, tofacitinib is FDA-approved for UC but not for CD. Mirikizumab is an approved drug for the treatment of ulcerative colitis (UC), and guselkumab is also approved for UC and is still under review for Crohn's disease (CD). Natalizumab is approved for CD but not for UC. Therefore, determining the diagnosis of UC or CD is crucial for more appropriate drug use, which is affected by drug reimbursement for specific disease subtypes.

[0044] In some embodiments, the methods and kits provided can be used for IBD subtype diagnosis, which is crucial for long-term management. For example, patients with UC and CD may require different long-term management strategies. UC management focuses on maintaining remission using medical therapies and regular monitoring to prevent sudden disease relapse and complications such as colorectal cancer. For CD, long-term management may involve a more proactive approach to prevent complications attributable to the risk of strictures and penetrating diseases, and patients require regular imaging and endoscopic evaluations. Furthermore, in subjects with severe or refractory disease requiring surgical intervention, a clear and accurate diagnosis is crucial for determining the type of surgery, which differs for CD and UC because the former is not curable by surgery. In some embodiments, the stratification for further treatment differs from that for CD and UC. CD is a more complex condition that can involve the upper gastrointestinal tract and lead to strictures and fistulas. In some embodiments, this requires more diagnostic procedures compared to UC, including computed tomography, ultrasound, magnetic resonance imaging, and upper gastrointestinal endoscopy to locate and assess affected areas in the small intestine.

[0045] In some embodiments, the novel methods and kits are useful for hospitals, institutions, or individuals willing to perform fecal / gut microbiota transplantation (FMT or IMT) for the treatment of IBD, CD, or UC.

[0046] In some embodiments, this disclosure provides compositions comprising a group of bacteria that are deficient in CD (also referred to in some embodiments as "CD-depleted species"), which may be used to treat CD patients. After determining that an individual has an increased risk of developing CD or already has CD, the individual may be treated with IMT or supplemented with the disclosed compositions (in the form of probiotic / microbe combination pills).

[0047] In some embodiments, this disclosure provides compositions comprising a group of bacteria that are deficient in UC (also referred to in some embodiments as “UC-depleted species”), which may be used to treat UC patients. After determining that an individual has an increased risk of developing UC or already has UC, the individual may be treated with IMT or supplemented with the disclosed compositions (in the form of probiotic / microbe combination pills). Attached Figure Description

[0048] Figure 1A The figure, according to an exemplary embodiment, shows the leading bacterial species associated with Crohn's disease (CD) identified in the study.

[0049] Figure 1B The figure, according to an exemplary embodiment, shows the association between CD, sex, age, and the relative abundance of 18 selected CD depletion and CD-enriched bacterial species.

[0050] Figure 2A-2C The figures, according to an exemplary embodiment, show the cohorts of 18 bacterial species markers identified by metagenomic sequencing in Hong Kong, China / HK, China. Figure 2A ), Hong Kong, China verification queue ( Figure 2B ) and the Australian (AUS) verification queue ( Figure 2C The relative abundance of CD in patients and normal individuals.

[0051] Figure 3A The figure, according to an exemplary embodiment, shows the receiver operating characteristic (ROC) curves and area under the curve (AUC) values ​​of a machine learning model (random forest model) using nine selected bacterial species markers identified by metagenomic sequencing in the training set, test set, Hong Kong validation cohort, and Australia (AUS) validation cohort.

[0052] Figure 3B The figure, according to an exemplary embodiment, shows the receiver operating characteristic (ROC) curves and area under the curve (AUC) values ​​of a machine learning model (random forest model) using nine selected bacterial species markers identified by metagenomic sequencing in three public datasets including subjects from the United States, the Netherlands, and China.

[0053] Figures 4A-4C The figure, according to an exemplary embodiment, shows the relative abundance (normalized abundance) of nine selected bacterial species markers identified by droplet digital PCR in CD patients and normal individuals (controls) in a discovery cohort subgroup.

[0054] Figure 5 The figure, according to an exemplary embodiment, shows the receiver operating characteristic (ROC) curve and area under the curve (AUC) value of a random forest model using nine selected bacterial species markers determined from the training and test sets via droplet digital PCR.

[0055] Figure 6A The stacked bar charts, according to an exemplary embodiment, compare the relative abundance (%) of bacterial species in starch degradation between healthy controls and CD patients.

[0056] Figure 6B The stacked bar charts, according to an exemplary embodiment, compare the relative abundance (%) of bacterial species in L-tryptophan biosynthesis between healthy controls and CD patients.

[0057] Figure 6C The stacked bar chart, according to an exemplary embodiment, compares the relative abundance (%) of bacterial species in the formation of thiamine pyrophosphate (yeast) from pyridine thiamine and hydroxy thiamine between healthy controls and CD patients.

[0058] Figure 6D The stacked bar charts, according to an exemplary embodiment, compare the relative abundance (%) of bacterial species in the superpathway of L-lysine, L-threonine, and L-methionine biosynthesis II between healthy controls and CD patients.

[0059] Figure 7 The figure, according to an exemplary embodiment, shows the correlation between functional ecological imbalance scores and disease probability generated by a machine learning model based on biomarkers of nine selected bacterial species.

[0060] Figures 8A-8C The figure, according to an exemplary embodiment, shows the relative abundance of nine selected bacterial species in healthy controls and patients in inactive and active states.

[0061] Figure 8D The figure, according to an exemplary embodiment, shows a comparison of disease probabilities between CD patients in an inactive state (N = 69) and an active state (N = 9), calculated using a random forest model.

[0062] Figure 8EThe figure, according to an exemplary embodiment, shows the receiver operating characteristic (ROC) curves and area under the curve (AUC) values ​​of a random forest model using nine selected bacterial species markers identified by metagenomics in CD patients and healthy controls during remission.

[0063] Figures 9A-9C The figure, according to an exemplary embodiment, shows the performance of the CD diagnostic model in distinguishing CD patients from other patients with non-IBD GI diseases.

[0064] Figures 10A-10B The figure illustrates the performance of the CD diagnostic model in distinguishing CD from other non-IBD and non-GI diseases, according to an exemplary embodiment.

[0065] Figure 10C The figure, according to an exemplary embodiment, shows the ROC curves for classifying CD patients with all other non-IBD (GI and nonGI) diseases using nine selected CD biomarkers.

[0066] Figure 11 The figure, according to an exemplary embodiment, shows the ROC curves for classifying CD patients from UC patients using nine selected CD biomarkers.

[0067] Figure 12 An exemplary method is shown, according to an exemplary embodiment, for determining the risk, diagnosing, preventing, or treating Crohn's disease (CD) in an individual.

[0068] Figure 13A The figure, according to an exemplary embodiment, shows the leading bacterial species associated with ulcerative colitis (UC) identified in the study.

[0069] Figure 13B The figure, according to an exemplary embodiment, shows the association between UC, sex, age and the relative abundance of 18 selected UC depletion and UC enrichment bacterial species calculated by the generalized linear model MaAsLin2.

[0070] Figures 14A-14B The figures, according to an exemplary embodiment, show the cohort of 18 bacterial species markers identified by metagenomic sequencing in Hong Kong, China. Figure 14A ) and Hong Kong, China verification queue ( Figure 14B The relative abundance of UC in UC patients and normal individuals.

[0071] Figure 15AThe figure, according to an exemplary embodiment, shows the receiver operating characteristic (ROC) curves and area under the curve (AUC) values ​​of a machine learning model (random forest model) using 10 selected bacterial species markers identified by metagenomic sequencing in the training set, test set, and Hong Kong validation cohort.

[0072] Figure 15B The figure, according to an exemplary embodiment, shows the receiver operating characteristic (ROC) curves and area under the curve (AUC) values ​​of a machine learning model (random forest model) using 10 selected bacterial species markers determined by metagenomic sequencing in three public datasets including subjects from the United States, the Netherlands, and China.

[0073] Figures 16A-16C The figure, according to an exemplary embodiment, shows the relative abundance (normalized abundance) of 10 selected bacterial species markers identified by droplet digital PCR in UC patients and normal individuals (controls) in a discovery cohort subgroup.

[0074] Figure 17 The figure, according to an exemplary embodiment, shows the receiver operating characteristic (ROC) curve and area under the curve (AUC) value of a random forest model using 10 selected biomarkers determined from the training and test sets via droplet digital PCR.

[0075] Figure 18A The stacked bar chart, according to an exemplary embodiment, compares the relative abundance (%) of selected bacterial species in L-arginine biosynthesis I (via L-ornithine) between healthy controls and UC patients.

[0076] Figure 18B The stacked bar chart, according to an exemplary embodiment, compares the relative abundance (%) of selected bacterial species in L-valine biosynthesis between healthy controls and UC patients.

[0077] Figure 18C The stacked bar chart, according to an exemplary embodiment, compares the relative abundance (%) of selected bacterial species in L-ornithine biosynthesis I between healthy controls and UC patients.

[0078] Figure 18D The stacked bar chart, according to an exemplary embodiment, compares the relative abundance (%) of selected bacterial species in de novo adenosine nucleotide biosynthesis between healthy controls and UC patients.

[0079] Figure 19 The figure, according to an exemplary embodiment, shows the correlation between functional ecological imbalance scores and disease probability generated by a machine learning model based on biomarkers of 10 selected bacterial species.

[0080] Figures 20A-20C The figure, according to an exemplary embodiment, shows the relative abundance of 10 selected bacterial species in healthy controls and patients in inactive and active states.

[0081] Figure 20D The figure, according to an exemplary embodiment, shows a comparison of disease probabilities between UC patients in an inactive state (N = 110) and an active state (N = 12) calculated using a random forest model.

[0082] Figure 20E The figure, according to an exemplary embodiment, shows the receiver operating characteristic (ROC) curves and area under the curve (AUC) values ​​of a random forest model using 10 selected bacterial species markers identified by metagenomics in UC patients and healthy controls during remission.

[0083] Figures 21A-21C The figure illustrates, according to an exemplary embodiment, the performance of the UC diagnostic model in distinguishing UC patients from other patients with non-IBD GI diseases.

[0084] Figures 22A-22B The figure illustrates the performance of the UC diagnostic model in distinguishing UC from other non-IBD, non-GI diseases, according to an exemplary embodiment.

[0085] Figure 22C The figure, according to an exemplary embodiment, shows the ROC curves that classify UC patients from those with all other non-IBD (GI and nonGI) diseases using 10 selected UC biomarkers.

[0086] Figure 23 An exemplary method is shown, according to an exemplary embodiment, for determining the risk, diagnosing, preventing, or treating ulcerative colitis (UC) in an individual.

[0087] Figure 24A The diagram shows an overview of the research workflow.

[0088] Figure 24B The image is a violin plot showing the Shannon index and observed species of the fecal microbiome in UC patients (N = 205), CD patients (N = 174), and controls (N = 118).

[0089] Figure 24C The principal coordinate analysis (PCoA) plot shows the different microbial compositions among the groups (174 CD patients, 205 UC patients, and 118 controls).

[0090] Figure 24C'The multivariate analysis plot shows the amount of explained variance and the corresponding p-values ​​for species-level Bray-Curtis dissimilarity determined by PERMANOVA.

[0091] Figure 24D The stacked bar chart shows the relative abundance of the six most abundant phyla in UC patients (N = 205), CD patients (N = 174), and controls (N = 118).

[0092] Figure 24D' This is a comparison chart of the relative abundance of the six gates in UC patients, CD patients, and healthy controls.

[0093] Figure 25C The figure shows the relative abundance of ten bacterial species biomarkers in the UC (N = 205) and control group (N = 118).

[0094] Figure 25D The figure shows the relative abundance of biomarkers for nine bacterial species in CD (N = 174) and control group (N = 118).

[0095] Figure 25E The figure shows the performance of the model with biomarkers of ten UC bacterial species in classifying UC.

[0096] Figure 25F The figure shows the performance of the model with biomarkers of nine CD bacterial species in classifying CDs.

[0097] Figure 25G The figure shows the Shapley Explanation (SHAP) values ​​for ten UC bacterial species biomarkers for each sample.

[0098] Figure 25H The figure shows the Shapley Explanation (SHAP) values ​​for the nine CD bacterial species biomarkers for each sample.

[0099] Figure 25K The figure shows the performance of models using different numbers of features to distinguish between UC patients and healthy controls.

[0100] Figure 25L The figure shows the performance of models using different numbers of features to distinguish CD patients from healthy controls.

[0101] Figures 26A-26D The figure shows the differential functional pathways between UC / CD patients and healthy controls and their correlation with bacterial species biomarkers.

[0102] Figure 26EThe figure shows the distribution of functional ecodysregulation scores between samples and controls, determined by median Bray-Curtis dissimilarity.

[0103] Figure 26F The figure shows a comparison of functional ecological disorder scores among UC (N = 205), CD (N = 174), and control (N = 118).

[0104] Figure 27A The figure shows the relative abundance of ten UC bacterial species biomarkers in healthy controls and UC patients in an inactive and active state.

[0105] Figure 27B The figure shows a comparison of disease probabilities between UC patients in an inactive state (N = 110) and an active state (N = 11), calculated using a random forest model.

[0106] Figure 27C The figure shows the model’s performance in distinguishing between inactive UC patients (N = 40) and healthy controls (N = 42).

[0107] Figure 27D The figure shows the relative abundance of nine CD bacterial species biomarkers in healthy controls and CD patients in an inactive and active state.

[0108] Figure 27E The figure shows a comparison of disease probabilities between CD patients in an inactive state (N = 69) and an active state (N = 9), calculated using a random forest model.

[0109] Figure 27F The figure shows the model’s performance in distinguishing between inactive CD patients (N = 69) and healthy controls (N = 108).

[0110] Figure 28A The figure shows the bacterial species biomarker characteristics for UC diagnosis in patients and healthy individuals used in the discovery cohort, validation cohort, and three downloaded public datasets.

[0111] Figure 28B The figure shows the bacterial species biomarker characteristics for CD diagnosis in patients and healthy individuals used in the discovery cohort, validation cohort, and three downloaded public datasets.

[0112] Figure 28C The figure shows the disease probabilities calculated using a random forest model between the discovery cohort in Hong Kong, China, the validation cohort from Hong Kong and Australia, and public datasets (US, Netherlands, and China) of UC / CD patients and healthy controls.

[0113] Figure 28D The figure shows the correlations among ten UC bacterial species markers.

[0114] Figure 28E The figure shows the correlations among the nine CD bacterial species markers.

[0115] Figure 29A The figure shows the performance of the model with ten selected bacterial species biomarkers for UC in a validation cohort in Hong Kong, China, in classifying UC patients versus controls.

[0116] Figures 29B-29C The figure shows the performance of the model with nine CD selected bacterial species biomarkers in a validation cohort in Hong Kong and Australia in classifying CD patients versus controls.

[0117] Figure 29D The figure shows the performance of the model with selected bacterial species biomarkers in classifying UC patients versus controls across three downloaded public datasets.

[0118] Figure 29E The figure shows the performance of the model with selected bacterial species biomarkers in classifying CD patients versus controls on three downloaded public datasets.

[0119] Figure 29F-29G The figure shows the associations between the relative abundance of disease group (UC or CD), geographic, ethnic, and bacterial species biomarkers, calculated using MaAsLin2 across all IBD cohorts.

[0120] Figure 29H The figure shows the performance of the model with selected bacterial species biomarkers in classifying UC patients (n = 817) versus controls (n = 1746) across all UC validation cohorts.

[0121] Figure 29I The figure shows the performance of the model with selected bacterial species biomarkers in classifying CD patients (n = 1065) versus controls (n = 1873) across all CD validation cohorts.

[0122] Figure 29J The figure shows the model performance in distinguishing between treated and treatment-naïve UC patients and controls in two downloaded public datasets.

[0123] Figure 29K The figure shows the model performance in distinguishing between treated and treatment-naïve CD patients and controls in two downloaded public datasets.

[0124] Figure 29LThe figure shows the model performance in distinguishing UC patients from controls on two downloaded public datasets compared to the fecal calprotectin test.

[0125] Figure 29M The figure shows the model performance in distinguishing CD patients from controls on two downloaded public datasets compared to the fecal calprotectin test.

[0126] Figure 30A The figure shows the relative abundance of 10 UC bacterial species biomarkers in UC and other non-IBD disease groups.

[0127] Figure 30B The figure shows the relative abundance of biomarkers of nine CD bacterial species in CD and other non-IBD disease groups.

[0128] Figure 31A The figure shows the composition of the international multi-disease dataset from different countries and regions.

[0129] Figure 31B The figure shows a comparison of the disease probabilities generated by UC models based on biomarkers of ten UC bacterial species in controls (n = 2391), CVD (n = 143), obesity (n = 318), CA (n = 230), CRC (n = 372), IBS-D (n = 146), and UC patients (n = 817).

[0130] Figure 31C The figure shows the performance of the UC model in classifying UC patients (n = 817) from other non-IBD subjects (n = 3600).

[0131] Figure 31D The figure shows a comparison of the disease probabilities generated by CD models based on biomarkers of nine CD bacterial species in controls (n = 2391), CVD (n = 143), obesity (n = 318), CA (n = 230), CRC (n = 372), IBS-D (n = 146), and CD patients (n = 1065).

[0132] Figure 31E The figure shows the performance of the CD model in classifying CD patients (n = 1065) from other non-IBD subjects (n = 3600).

[0133] Figure 32A The figure shows the ROC of the general IBD model in classifying IBD from control and non-IBD in the test set, IBD validation cohort and non-IBD cohort.

[0134] Figure 32B The figure shows a typical criteria (STARD) diagram for reporting diagnostic accuracy studies, which reports the participant flow in the Independent International IBD Cohort (IBD = 1882, Control = 2027).

[0135] Figure 32C The figure shows a comparison of the diagnostic performance of the universal IBD model and fecal calprotectin in classifying IBD and IBS subjects.

[0136] Figure 33A The figure shows the group design for multiplex droplet digital PCR for UC and CD bacterial species markers.

[0137] Figure 33B The figure shows the correlation between the abundance of biomarkers for ten UC bacterial species, determined by metagenomics and multiplex droplet digital PCR methods.

[0138] Figure 33C The figure shows the correlation between the abundance of biomarkers for nine CD bacterial species, determined by metagenomics and multiplex droplet digital PCR methods.

[0139] Figure 34A The figure shows the relative abundance of ten bacterial biomarkers in the discovery cohort (205 UCs; 84 controls) and the healthy control group.

[0140] Figure 34B The figure shows the diagnostic performance of ten bacterial species biomarkers identified by m-ddPCR in distinguishing UC patients from healthy controls in the discovery cohort (test set, N = 62) and the Hong Kong, China cohort (N = 108).

[0141] Figure 34C The figure shows the relative abundance of nine bacterial biomarkers in the CD cohort and the healthy control group in the discovery cohort (172 CDs; 86 controls).

[0142] Figure 34D The figure shows the diagnostic performance of nine bacterial species biomarkers identified by m-ddPCR in distinguishing CD patients from healthy controls in the discovery cohort (test set, N = 66), the Hong Kong cohort (N = 153), and the Australian cohort (N = 177).

[0143] Figure 34E The figure shows the diagnostic performance of the fecal calprotectin test and the UC model with biomarkers of ten bacterial species identified by m-ddPCR in the Canadian cohort (100 UCs, 53 controls) and the Taiwan cohort (40 UCs, 40 controls).

[0144] Figure 34F The figure shows the diagnostic performance of the fecal calprotectin test and CD models with biomarkers of ten bacterial species identified by m-ddPCR in a Canadian cohort (100 CDs, 53 controls) and a Taiwan cohort (40 CDs, 40 controls).

[0145] Figure 34G The figure shows a comparison of disease probabilities calculated by the UC / CD model using m-ddPCR data and fecal calprotectin results between inactive and active UC / CD patients and healthy controls in the Canadian and Taiwan cohorts.

[0146] Figure 34H The figure shows the diagnostic performance of the UC model in distinguishing between inactive UC patients (N = 81) and healthy controls (N = 93) using fecal calprotectin testing and ten bacterial biomarkers identified by m-ddPCR.

[0147] Figure 35A The figure shows the difference in the probability of disease (POD) in UC calculated by the metagenomics-based model and the m-ddPCR-based model.

[0148] Figure 35B The figure shows the difference in the probability of disease (POD) in CD calculated by the metagenomics-based model and the m-ddPCR-based model.

[0149] Figure 36 The diagram illustrates the workflow of the testing algorithm using the IBD model and the IBD subtype model, according to an exemplary embodiment. Detailed Implementation

[0150] As used herein and in the claims, the terms “comprising” (or any related form, such as “comprise” and “comprises”), “including” (or any related form, such as “include” or “includes”), “containing” (or any related form, such as “contain” or “contains”) mean to include the following elements but not exclude other elements. It should be understood that for each embodiment in which the terms “comprising” (or any related form, such as “comprise” and “comprises”), “including” (or any related form, such as “include” or “includes”), or “containing” (or any related form, such as “contain” or “contains”) are used, this disclosure / application also includes alternative embodiments in which the terms “comprising”, “including”, or “contains” are replaced with “consistently consisting of” or “consisting of”. These alternative embodiments using the terms "composed of" or "substantially composed of" are understood to be smaller-scale embodiments that "comprise", "include", or "contain" the embodiments.

[0151] For example, alternative embodiments of "a composition comprising A, B, and C" would be "a composition consisting of A, B, and C" and "a composition consisting essentially of A, B, and C". Even if the latter two embodiments are not explicitly stated, this disclosure / application includes those embodiments. Furthermore, it should be understood that the scope of the three embodiments listed above is different.

[0152] For clarity, “comprising,” “including,” and “containing,” as well as any related forms, are open-ended terms that allow for additional elements or features beyond the specified mandatory elements, while “composes of” is a closed-ended term that is limited to the elements listed in the claims and excludes any elements, steps, or ingredients not specified in the claims.

[0153] For clarity, "characterized as" or "characterized in" (along with their related forms as described above) does not limit or change the nature of whether the following list of terms is open or closed. For example, in a claim relating to "a composition comprising A, B, C and characterized in D, E, and F," elements D, E, and F remain open-ended terms, and the claim is intended to include other elements due to the use of the word "comprising" preceding the claim.

[0154] The phrase "consistently of..." limits the scope of the claim to specific materials, components, or steps ("basic elements") that do not materially affect one or more essential features of the claimed invention. In some embodiments, essential features are one or more basic and novel features of the claimed invention. For example, in some embodiments, the basic element of the composition disclosed herein may be compound A in the form of "X mg to Y mg". Even if the composition contains additional excipients, such "consistently of compound A" embodiments still include compositions having the aforementioned additional excipients, provided that the additional excipients do not materially affect the essential features of the compound, such as the ability of compound A to bind to target XX or treat disease YY.

[0155] As used herein, the singular forms “a / an” and “the” are intended to also include the plural forms unless the context clearly indicates otherwise. When a range is mentioned in the specification, the range is understood to include each discrete point within the range. For example, 1 to 7 means 1, 2, 3, 4, 5, 6, and 7.

[0156] As used herein and in the claims, an "effective amount" is an amount that effectively achieves the desired effect of at least a measurable amount. In some embodiments, the amount can effectively increase amino acid biosynthesis, carbohydrate degradation, cofactors, carriers, and / or vitamin biosynthesis in cells. In some embodiments, providing an effective amount includes, but is not limited to, administering a therapeutic agent to a patient (e.g., orally, intravenously, subcutaneously, topically, via inhalation or nasal administration, or intramuscularly), or contacting cells with the composition, whether in vitro, ex vivo, or in vivo.

[0157] As used herein, the term "composition" refers to a formulation containing one or more active ingredients. In some instances, the composition comprises one or more bacterial species and optionally an acceptable carrier. In some instances, the composition serves as a medicine capable of treating CD, UC, and / or IBD in subjects of need. In some embodiments, the acceptable carrier is a pharmaceutically acceptable carrier. In some embodiments, the carrier is a dietary formulation, food, or beverage. In some instances, the composition is used as a nutritional supplement, dietary formulation, functional food, medical food, food, or beverage for use in maintaining general health in subjects. In some instances, the composition is in the form of a probiotic / microbial combination pill.

[0158] As used herein and in the claims, the terms “prevent,” “preventing,” “preventive / preventative,” or “prevention” refer to a method of reducing the risk of the onset, recurrence, or spread of a disease or disorder or one or more symptoms thereof.

[0159] As used herein, the terms “treat,” “treating,” or “treatment” refer to the following methods of: relieving, eliminating, or improving symptoms of a disease or condition in a preventive and / or therapeutic manner; preventing additional symptoms; improving or preventing the basal metabolic cause of symptoms; inhibiting a disease or condition; halting the progression of a disease or condition; alleviating a disease or condition; causing the remission of a disease or condition; reducing the condition caused by a disease or condition; or terminating the symptoms of a disease or condition.

[0160] As used herein, the term “sequencing” refers to techniques that involve obtaining sequence information from one or more nucleic acid molecules by determining the identity and / or sequence of at least some nucleotides within each nucleic acid molecule.

[0161] As used herein, the term "relative abundance," when used in the context of describing the presence of a particular bacterial species relative to all bacterial species present in the same sample (e.g., in an individual's fecal sample), refers to the relative quantity of said bacterial species relative to all bacterial species in said particular sample. For example, the relative abundance of a particular bacterial species can be determined by comparing the amount of species-specific DNA in a given sample (e.g., determined by quantitative polymerase chain reaction) with the amount of bacterial DNA in all bacteria in the same sample (e.g., determined by quantitative polymerase chain reaction (PCR) and / or sequencing based on 16S rRNA sequences).

[0162] As used herein, the term "amplification" or "amplify" includes methods for replicating a target nucleic acid sequence, thereby increasing the copy number of the target nucleic acid sequence. Amplification can be exponential or linear, for example, by polymerase chain reaction (PCR), real-time PCR, droplet digital PCR, or other nucleic acid amplification methods known in the art. Sequences amplified in this manner form "amplification products".

[0163] As used herein, the term "primer set" refers to a pair of primers both targeting a target nucleic acid sequence. In some embodiments, a primer set targeting a particular target nucleic acid sequence contains a forward primer and a reverse primer, each of which hybridizes to a different strand (sense or antisense) of the target nucleic acid sequence under appropriate conditions.

[0164] As used herein, the term "probe" refers to a nucleic acid molecule attached to at least one detectable label, wherein the nucleic acid molecule comprises a probe sequence capable of hybridizing with at least a portion of a target nucleic acid sequence. Examples of detectable labels include, but are not limited to, radioactive isotopes, enzyme substrates, cofactors, ligands, chemiluminescent or fluorescent agents, haptens, and enzymes.

[0165] As used herein, the term "fluorophore-quencher probe" refers to a nucleic acid molecule labeled with a fluorophore and a quencher, wherein the nucleic acid molecule contains a probe sequence capable of hybridizing with at least a portion of a target nucleic acid sequence. In some embodiments, the fluorophore is attached to the 5' end of the nucleic acid molecule, and the quencher is attached to the 3' end of the nucleic acid molecule. During PCR amplification, if the fluorophore-quencher probe binds to the target nucleic acid sequence, the fluorophore dissociates from the quencher, resulting in the release of detectable and quantifiable fluorescence.

[0166] As used herein, the terms “marker,” “bacterial marker,” “bacterial species marker,” “bacterial biomarker,” or “bacterial species biomarker” refer to a preselected bacterial species and / or gene sequence used in a method / kit as described herein. For example, the relative abundance of deoxyribonucleic acid (DNA) fragments of each bacterial biomarker in a fecal sample from each subject can be measured.

[0167] As used herein, the term "reference dataset" refers to a dataset generated from a cohort of subjects with a confirmed diagnosis of CD, UC, and / or IBD and another cohort of subjects not diagnosed with CD, UC, and / or IBD and not exhibiting any signs or symptoms of CD, UC, and / or IBD or gastrointestinal disease. It can be generated by measuring the relative abundance of each "bacterial marker" for each subject in said cohort. The diagnosis of subjects in said cohort can be determined by standard methods. For example, a diagnosis can be made according to international guidelines through a combination of clinical, biochemical, fecal, endoscopic, cross-sectional imaging, and histological studies. For a "reference cohort" to be properly established, the control group must include a sufficient number of individuals without CD, UC, and / or IBD (e.g., at least 10, 12, 15, 20, 24, or more individuals) to provide samples for determining one or more average levels of one or more preselected bacterial species.

[0168] As used in this article, the terms “normal individual,” “healthy control,” or “control” refer to an individual who does not have a serious disease or pre-existing bowel disorder (such as inflammatory bowel disease, cancer, advanced adenoma, irritable bowel syndrome, or other GI symptoms).

[0169] As used in this article, the term “individual who does not have IBD” refers to an individual who does not have IBD but may have other diseases or conditions such as gastrointestinal diseases or conditions, irritable bowel syndrome, obesity, acid reflux, colonic polyps, functional dyspepsia, gastrointestinal stromal tumors (GIST) or other GI symptoms.

[0170] As used herein, the term "active IBD" refers to an IBD patient with severe IBD-related symptoms, while the term "inactive IBD" refers to an IBD patient with mild to moderate IBD-related symptoms. Similarly, the term "active UC" refers to an UC patient with severe UC-related symptoms, while the term "inactive UC" refers to an UC patient with mild to moderate UC-related symptoms; the term "active CD" refers to a CD patient with severe CD-related symptoms, while the term "inactive CD" refers to a CD patient with mild to moderate CD-related symptoms.

[0171] As used herein, the term "training dataset" refers to a subset of the "reference dataset." It includes the known health classifications of the subjects (CD, UC, and / or IBD or controls) and the feature vectors of the subjects, which can be used to train machine learning models.

[0172] As used in this article, “tuning” refers to the process of adjusting the parameters or hyperparameters of a machine learning algorithm or model to optimize its performance on a given task or dataset. In some instances, during the tuning process, different values ​​or combinations of parameters are explored and evaluated to find settings that produce the best results. The goal is to fine-tune the model’s performance by finding the optimal configuration that minimizes error, maximizes accuracy, or achieves the desired results.

[0173] As used herein, the terms “active” or “inactive” refer to alternation of active phases of UC, CD, and / or IBD. In some embodiments, UC is characterized by alternation of active phases, including flare-ups (i.e., active UC) and remission (i.e., inactive UC). In some embodiments, CD is characterized by alternation of active phases, including flare-ups (i.e., active CD) and remission (i.e., inactive CD). In some embodiments, disease activity in UC patients is assessed using the Mayo Criterion, where patients are categorized as having active disease or having inactive disease; disease activity in CD patients is assessed using the Crohn's Disease Activity Index or the Harvey-Bradshaw Index, where patients are categorized as having active disease or having inactive disease. In some embodiments, the term “active IBD” refers to active UC or active CD. In some embodiments, the term “inactive IBD” refers to inactive UC or inactive CD.

[0174] When this document describes “treating an individual, respectively, with appropriate IBD treatment, UC treatment, or CD treatment if it is determined that the individual has an increased risk of developing IBD, UC, or CD, or if the individual already has IBD, UC, or CD,” the term “respectively” means that if it is determined that the individual has an increased risk of developing a disease (e.g., UC), or if the individual already has a disease (e.g., UC), the individual is treated with an appropriate treatment for that disease (UC treatment). In some embodiments, there may be overlap between treatments effective for UC and another disease (e.g., CD, IBD, or other diseases). In some embodiments, the appropriate disease treatment is a treatment that has been regulatory approved for the disease. In some embodiments, the appropriate disease treatment is specific to the one disease (e.g., UC) and is not approved for other diseases (e.g., CD). In some embodiments, the treatment is approved for UC and another disease, such as CD.

[0175] When this article describes an individual as “not receiving XX treatment” (where XX is CD, UC, or IBD), it means that the individual is not currently receiving XX treatment, or that the individual has not recently received XX treatment (e.g., not receiving XX treatment 1–3x months prior to sample collection).

[0176] When this article describes an individual as “treatment-naïve” (where XX is CD, UC, or IBD), it means that the individual had never received treatment with XX prior to sample collection.

[0177] When this document describes an individual as “exposed to XX treatment” (where XX is CD, UC, or IBD), it means that the individual is currently undergoing XX treatment, or the individual recently received XX treatment (e.g., received XX treatment 1–3 months prior to sample collection), or the individual received XX treatment at any time prior to sample collection.

[0178] Although specific embodiments are mentioned in the description, this disclosure should not be construed as limiting it to the embodiments set forth herein.

[0179] This document provides examples of certain embodiments of this disclosure that are described in more detail. The examples provided herein are for illustrative purposes only and are not intended to limit the scope of the invention in any way. All references set forth below and elsewhere in this application are hereby incorporated by reference. Example 1

[0180] In some embodiments, a method is provided for determining, diagnosing, preventing, or treating Crohn's disease (CD) in an individual, the method comprising the steps of: (a) determining the relative abundance of a set of bacterial species in a fecal sample from the individual, wherein the set of bacterial species comprises one or more bacterial species selected from the group consisting of: *Dolphus formateriate*, *Eubacterium pickytiae*, *Roseobacterium inulinum*, *Roseobacter rumenella*, *Broutella ovalis*, *Eubacterium truncatella*, *Anaerobic bacillus hardruciformis*, *Lawsonia suis*, *Oscillatoria* species CAG:241, *Oscillatoria* species 57_20, *Lactobacillus rhamnosus*, *Eubacterium* species CAG: 274, *Romebuts ileum*, *Bacteroides fragilis*, *Escherichia coli*, *Eubacterium gingivalis*, *Actinomyces* species oral taxa 181, and *Actinomyces* species S6-Spd3; (b) comparing the relative abundance of each bacterial species in the bacterial species set from the individual with the relative abundance of each bacterial species in the bacterial species set in the reference dataset to generate a risk score; (c) if the risk score is higher than a cutoff value, determining that the individual has an increased risk of having CD or is currently suffering from CD; and (d) if the individual is determined to have an increased risk of having CD or is currently suffering from CD, optionally treating the individual.

[0181] In some embodiments, the bacterial species set includes one or more bacterial species selected from the group consisting of: *Gimidae formic acid*, *Eubacterium hominis*, *Broutella ovalis*, *Roseobacterium inulinum*, *Clostridium sacchariformis*, *Eubacterium rectum*, *Roseobacterium falciparum*, *Bacillus crypticarum*, *Collinsella aerogenes*, *Bacillus preibryophyte*, *Anaerobic bacillus hardruciformis*, *Trichophyton molasses*, *Ruminococcus tortifolius*, *Clostridium truncatella*, *Bacteroides falciparum*, *Ruminococcus bryophyte*, *Roseobacterium entericarum*, *Adlerococcus brevis*, *Alternaria putrefactive*, *Eubacterium* species CAG: 38, *Roseobacter humanis*, *Agastache rugosa*, *Dolphobic formic acid*, *Bacteroides fecalis*, *Ostomyces viscerans*, *Ostomyces spp.* species 57-20, *Cercospora fecalith*, *Bacteroides var. ... 241. *Eubacterium filamentosa*, *Alternaria fenestrate*, *Bacteroides foetida*, *Ackermania spp.*, *Concomitant cocci*, *Toxic butyrosporum*, *Alternaria inconspicuous*, *Bacteroides monomorpha*, *Bacteroides foetida*, *Eubacterium* species CAG: 274. *Desulfovibrio suis*, *Bacteroides masei*, *Clostridium* species CAG: 299. *Eubacterium spp.*, *Romebuts ileum*, *Eubacterium spp.*, *Anaerobic bacillus hardrochetes*, *Lactobacillus rhussi*, *Ruminococcus ovalis*, *Hemolytic twins*, *Streptococcus leucis*, *Actinomyces cariogenicus*, *Actinomyces* species S6-Spd3, *Actinomyces griseus*, *Lactobacillus mucosae*, *Streptococcus parahaemolyticus*, *R. mucosae*, *Actinomyces* species oral taxonomy group 181, *Veillonella septum*, *Escherichia coli*, *Bacteroides fragilis*, and *Eubacterium gingivalis*.

[0182] In some embodiments, the reference dataset is obtained by determining the relative abundance of each bacterial species in a set of bacterial species from a subject group that includes CD patients and normal individuals in a reference cohort.

[0183] In some embodiments, a machine learning model is used to perform the comparison step in step (b) to generate a risk score.

[0184] In some embodiments, the machine learning model is constructed by: (i) obtaining a reference dataset by determining the relative abundance of each bacterial species in a bacterial species set from fecal samples of a subject group including CD patients and normal individuals in a reference cohort; (ii) dividing the reference dataset into a training dataset and a test dataset; (iii) dividing the training dataset into a training subset and a validation subset; (iv) training a candidate model using the training subset and evaluating the performance of the candidate model using the validation subset; (v) optionally tuning the candidate model by adjusting the hyperparameters in the candidate model; (vi) repeating steps (iv) and optionally (v) to generate a set of candidate models and selecting the best candidate model from the set of candidate models; (vii) evaluating the performance of the best candidate model using the test dataset; (viii) optionally training a final model with the reference dataset using a set of hyperparameters corresponding to the best candidate model to generate a final model; and (ix) using the best candidate model of step (vii), or if the final model was trained using step (viii), using the final model of step (viii) as the model for step (b). The machine learning model in the comparison step is used to generate the risk score. For clarity, the reference dataset used in step (viii) is a complete reference dataset, which includes both the training dataset and the test dataset.

[0185] In some embodiments, the machine learning model is a random forest model, and the risk score in step (b) is generated by the following steps: (b1) generating a set of decision trees using the relative abundance of the bacterial species set in the reference dataset via the random forest model; and (b2) running the relative abundance of the bacterial species set from individuals along the set of decision trees to generate a risk score.

[0186] In some embodiments, the best candidate model is the candidate model with the highest performance in the candidate model set.

[0187] In some embodiments, the performance of a candidate model is evaluated by the area under the curve (AUC) of the candidate model's receiver operating characteristic (ROC) curve. In some embodiments, the area under the curve (AUC) of the receiver operating characteristic (ROC) is also referred to as "AUROC".

[0188] In some embodiments, the cutoff value is in the range of 0.4 to 0.7.

[0189] In some embodiments, the cutoff value is 0.5.

[0190] In some embodiments, the cutoff value is determined by the following steps: (c1) calculating the Yoden index using the risk scores in the reference dataset; and (c2) determining the cutoff value based on the Yoden index.

[0191] In some embodiments, the bacterial species set includes one or more combinations of these bacterial species, wherein the combinations are selected from the group consisting of Tables 1.4 and 1.6.

[0192] In some embodiments, the bacterial species set consists essentially of: Actinomycete species oral taxa 181, Lawsonia non-saccharitroph, and Eubacterium species CAG: 274.

[0193] In some embodiments, the bacterial species set includes *Ruminococcus ovalis* (Broutella ovalis), *Lawsonia ovalis* (unsaccharitroph), *Roseidobacterium inulinum*, and optionally, *Actinomyces* species of oral taxa 181 and / or *Escherichia coli*.

[0194] In some embodiments, the bacterial species set includes Actinomycetes species oral taxa 181, Bacteroides fragilis, Escherichia coli, Lawsonia non-saccharitroph, Eubacterium species CAG: 274, Rosetteia inulinica, Rosetteia ovalis, Ruminococcus ovalis, and Dorgellella formica.

[0195] In some embodiments, the bacterial species set comprises the following: Actinomycete species oral taxa 181, Bacteroides fragilis, Escherichia coli, Lawsonia non-saccharitroph, Eubacterium species CAG: 274, Rosette of inulin, Rosette of enterica, Ruminococcus ovalis, and Dorperium formic acid-producing.

[0196] In some embodiments, the bacterial species set includes Bacteroides fragilis, Escherichia coli, Rosette of inulin, and Ruminococcus ovalis (Brutella ovalis), thereby enabling the method to identify an individual's risk of developing CD compared to other IBD subtypes or non-IBD diseases.

[0197] In some embodiments, other IBD subtypes include ulcerative colitis (UC), and non-IBD diseases include colorectal cancer, colorectal adenoma, irritable bowel syndrome (diarrhea subtype), obesity, and cardiovascular disease.

[0198] In some embodiments, the relative abundance of the bacterial species set is determined by metagenomic sequencing or droplet digital PCR (ddPCR).

[0199] In some embodiments, the relative abundance of the bacterial species set is determined by metagenomic sequencing, and the bacterial species set comprises a combination of one or more bacterial species selected from Table 1.4.

[0200] In some embodiments, the relative abundance of the bacterial species set is determined by ddPCR, and the bacterial species set comprises a combination of one or more bacterial species selected from Table 1.6.

[0201] In some embodiments, ddPCR is performed using one or more primer sets selected from Table 1.1.

[0202] In some embodiments, if an individual is determined to have an increased risk of developing CD or already has CD, the individual is treated with an appropriate CD treatment selected from the group consisting of: therapeutic treatment, fecal microbiota transplantation (FMT), intra-intestinal microbiota transplantation (IMT), lifestyle and dietary changes, and surgical intervention. In some embodiments, the therapeutic treatment is selected from the group consisting of: 5-aminosalicylate (5-ASA, such as mesalazine, sulfasalazine), steroids (such as budesonide, prednisolone), thiopurines (such as azathioprine, mercaptopurine, methotrexate), TNF inhibitors (such as infliximab, adalimumab, and sertozumab), anti-integrins (such as vedolizumab), and anti-interleukin (IL) 12 / 23 (such as ustekinumab).

[0203] In some embodiments, 5-aminosalicylate is mesalazine or sulfasalazine; steroid is budesonide or prednisolone; thiopurine is azathioprine, mercaptopurine, or methotrexate; TNF inhibitor is infliximab, adalimumab, or cetuzumab; anti-integrin is vedolizumab; and anti-interleukin (IL) 12 / 23 is ustekinumab.

[0204] In some embodiments, CD is in a remission period.

[0205] In some embodiments, a method is provided for identifying a fecal sample with an altered Crohn's disease microbiome, the method comprising the steps of: (a) determining the relative abundance of a set of bacterial species in the fecal sample, wherein the set of bacterial species comprises one or more bacterial species selected from the group consisting of: *Dolphobic formate*, *Eubacterium pickyii*, *Roseinia inulinica*, *Roseinia enterica*, *Ruminococcus ovalis* (*Broutella ovalis*), *Eubacterium truncatella*, *Anaerobic bacillus hardruciformis*, *Lawsonia solani*, *Oscillatoria* species CAG: 241, *Oscillatoria* species 57_20, *Lactobacillus rhamnosus*, *Eubacterium* species CAG: 274, *Romebuts ileum*, *Bacteroides fragilis*, *Escherichia coli*, *Eubacterium gingivalis*, *Actinomyces* species oral taxa 181, and *Actinomyces* species S6-Spd3; (b) The relative abundance of each bacterial species in the bacterial species set from the individual is compared with the relative abundance of each bacterial species in the bacterial species set in the reference dataset to generate a risk score; and (c) if the risk score is higher than a cutoff value, the fecal sample is determined to have an altered CD microbiome.

[0206] In some embodiments, a machine learning model is used to perform the comparison step in step (b) to generate a risk score.

[0207] In some embodiments, the machine learning model is constructed by: (i) obtaining a reference dataset by determining the relative abundance of each bacterial species in a bacterial species set from fecal samples of a subject group including CD patients and normal individuals in a reference cohort; (ii) dividing the reference dataset into a training dataset and a test dataset; (iii) dividing the training dataset into a training subset and a validation subset; (iv) training a candidate model using the training subset and evaluating the performance of the candidate model using the validation subset; (v) optionally tuning the candidate model by adjusting the hyperparameters in the candidate model; (vi) repeating steps (iv) and optionally (v) to generate a set of candidate models and selecting the best candidate model from the set of candidate models; (vii) evaluating the performance of the best candidate model using the test dataset; (viii) optionally training a final model with the reference dataset using a set of hyperparameters corresponding to the best candidate model to generate a final model; and (ix) using the best candidate model from step (vii), or if the final model was trained using step (viii), using the final model from step (viii) as the model for step (b). A machine learning model is used for comparison steps to generate risk scores.

[0208] In some embodiments, the machine learning model is a random forest model, and the risk score in step (b) is generated by the following steps: (b1) generating a set of decision trees using the relative abundance of the bacterial species set in the reference dataset via the random forest model; and (b2) running the relative abundance of the bacterial species set from individuals along the set of decision trees to generate a risk score. In some embodiments, the risk score is generated by averaging predictions across decision trees or by majority voting on the predicted class labels across decision trees, wherein the proportion of votes in each class across the set is the predicted probability (i.e., the risk score).

[0209] In some embodiments, the best candidate model is the candidate model with the highest performance in the candidate model set.

[0210] In some embodiments, the performance of a candidate model is evaluated by the area under the curve (AUC) of the receiver operating characteristic (ROC) curve of the candidate model.

[0211] In some embodiments, the cutoff value is in the range of 0.4 to 0.7.

[0212] In some embodiments, the cutoff value is 0.5.

[0213] In some embodiments, the cutoff value is determined by the following steps: (c1) calculating the Yoden index using the risk scores in the reference dataset; and (c2) determining the cutoff value based on the Yoden index.

[0214] In some embodiments, a method is provided in which the bacterial species set includes one or more combinations of such bacterial species, wherein the combinations are selected from the group consisting of Tables 1.4 and 1.6.

[0215] In some embodiments, the bacterial species set includes oral taxa of the genus *Actinomyces* 181, *Lawsonia unsacchariformis*, and *CAG: 274* of the genus *Eubacterium*. In some embodiments, the bacterial species set is substantially composed of oral taxa of the genus *Actinomyces* 181, *Lawsonia unsacchariformis*, and *CAG: 274* of the genus *Eubacterium*.

[0216] In some embodiments, the bacterial species set includes *Brutella ovalis*, *Lawsonia ovalis*, *Roseidon*, and optionally *Actinomyces* species oral taxonomy group 181 and / or *Escherichia coli*. In some embodiments, the bacterial species set comprises: *Brutella ovalis*, *Lawsonia ovalis*, and *Roseidon*. In some embodiments, the bacterial species set comprises: *Brutella ovalis*, *Lawsonia ovalis*, *Roseidon*, *Actinomyces* species oral taxonomy group 181, and *Escherichia coli*. In some embodiments, the bacterial species set comprises: *Brutella ovalis*, *Lawsonia ovalis*, *Roseidon*, and bacterial species selected from the group consisting of: *Actinomyces* species oral taxonomy group 181 and *Escherichia coli*.

[0217] In some embodiments, the bacterial species set includes Actinomycetes species oral taxa 181, Bacteroides fragilis, Escherichia coli, Lawsonia non-saccharitroph, Eubacterium species CAG: 274, Rosetteia inulinica, Rosetteia ovalis, Ruminococcus ovalis, and Dorgellella formica.

[0218] In some embodiments, the bacterial species set comprises the following: Actinomycete species oral taxa 181, Bacteroides fragilis, Escherichia coli, Lawsonia non-saccharitroph, Eubacterium species CAG: 274, Rosette of inulin, Rosette of enterica, Ruminococcus ovalis, and Dorperium formic acid-producing.

[0219] In some embodiments, the relative abundance of the bacterial species set is determined by metagenomic sequencing or droplet digital PCR (ddPCR).

[0220] In some embodiments, the relative abundance of the bacterial species set is determined by metagenomic sequencing, and the bacterial species set comprises a combination of one or more bacterial species selected from Table 1.4.

[0221] In some embodiments, the relative abundance of the bacterial species set is determined by ddPCR, and the bacterial species set comprises a combination of one or more bacterial species selected from Table 1.6.

[0222] In some embodiments, ddPCR is performed using one or more primer sets selected from Table 1.1.

[0223] In some embodiments, a method is provided to increase amino acid biosynthesis in an individual, wherein the amino acid is L-arginine, L-ornithine, or L-valine, the method comprising providing the individual with an effective amount of one or more bacterial species selected from the group consisting of: ovomycosis (Broutella ovalis), inulin-eating Rosette, formic acid-producing Dorperella, Eubacterium species CAG: 274, and Enterobacter Rosette.

[0224] In some embodiments, a method is provided to increase amino acid biosynthesis in cells, wherein the amino acid is L-arginine, L-ornithine, or L-valine, the method comprising providing the cells with an effective amount of one or more bacterial species selected from the group consisting of: ovale-forming rumenococci (ovale-forming broutitidis), inulin-eating Rosette, formic acid-producing Dorperella, Eubacterium species CAG: 274, and Enterobacter Rosette.

[0225] In some embodiments, a method for increasing carbohydrate degradation in an individual is provided, the method comprising providing the individual with an effective amount of one or more bacterial species selected from the group consisting of: ovale-forming rumenococci (ovale-forming Broutella), inulin-eating Rosette, formic acid-producing Dorperella, Eubacterium species CAG: 274, and Enterobacter Rosette.

[0226] In some embodiments, a method for increasing carbohydrate degradation in cells is provided, the method comprising providing the cells with an effective amount of one or more bacterial species selected from the group consisting of: ovale-forming rumenococci (ovale-forming Brontë), inulin-eating Rosette, formic acid-producing Dorper, Eubacterium species CAG: 274, and Enterobacter Rosette.

[0227] In some embodiments, a method is provided to increase the biosynthesis of cofactors, carriers, and vitamins in an individual, wherein the vitamin is thiamine pyrophosphate, the method comprising providing the individual with an effective amount of one or more bacterial species selected from the group consisting of: ovomycosis (Broutella ovalis), inulin-eating Rosette, formic acid-producing Dorperella, Eubacterium species CAG: 274, and Enterobacter Rosette.

[0228] In some embodiments, a method is provided for increasing the biosynthesis of cofactors, carriers, and vitamins in cells, wherein the vitamin is thiamine pyrophosphate, the method comprising providing the cells with an effective amount of one or more bacterial species selected from the group consisting of: ovale-forming rumenococci (ovale-forming Broutella), inulin-eating Rosette, formic acid-producing Dorperella, Eubacterium species CAG: 274, and Enterobacter Rosette.

[0229] In some embodiments, a method is provided to increase amino acid biosynthesis in an individual, wherein the amino acid is L-tryptophan, the method comprising providing the individual with an effective amount of ovoid rumenococci (ovoid broutella), inulin-eating Rosette, formic acid-producing Dorperella, Eubacterium species CAG: 274, or Enterobacter Rosette.

[0230] In some embodiments, a method is provided to increase amino acid biosynthesis in cells, wherein the amino acid is L-tryptophan, the method comprising providing the cells with an effective amount of ovoid rumenococci (ovoid broutella), inulin-eating Rosette, formic acid-producing Dorperella, Eubacterium species CAG: 274, or Enterobacter Rosette.

[0231] In some embodiments, the cell is part of a living organism.

[0232] In some embodiments, a composition is provided comprising Lawsonia non-saccharitroph and Eubacterium species CAG: 274.

[0233] In some embodiments, a composition is provided comprising ovoid rumenococci (ovoid brouteria), sugar-insoluble Lawsonia, and inulin-eating Rosette.

[0234] In some embodiments, a composition as described herein is provided, further comprising one or more bacterial species selected from the group consisting of: formic acid-producing Dorperella, picky bacteria, enteric Rosella, protruding bacteria, anaerobic bacillus, vibrio species CAG: 241, vibrio species 57_20, Lactobacillus rogueside, Eubacterium species CAG: 274, and Rombutzella ileum.

[0235] In some embodiments, a dietary composition is provided comprising the composition described in any of the foregoing aspects or embodiments, for example, wherein the dietary composition is selected from medical foods, functional foods, or supplements. In some embodiments, a nutritional supplement, dietary formulation, functional food, medical food, food, or beverage is provided comprising the composition described herein for use in a subject to maintain general health. Another embodiment provides a nutritional supplement, dietary formulation, functional food, medical food, food, or beverage comprising the composition described herein for use in a subject to manage CD.

[0236] In some embodiments, a kit is provided for determining the risk of or diagnosing Crohn's disease (CD) in an individual, the kit comprising reagents for detecting a set of bacterial species selected from the group consisting of: *Dolphus formateriate*, *Eubacterium pickyii*, *Roseobacterium inulinum*, *Roseobacter rumenella*, *Eubacterium ovale*, *Eubacterium truncatella*, *Anaerobic bacillus hardruciformis*, *Lawsonia suis*, *Oscillatoria* species CAG: 241, *Oscillatoria* species 57-20, *Lactobacillus rhamnosus*, *Eubacterium* species CAG: 274, *Romebuts ileum*, *Bacteroides fragilis*, *Escherichia coli*, *Eubacterium gingivalis*, *Actinomyces* species oral taxonomy group 181, and *Actinomyces* species S6-Spd3.

[0237] In some embodiments, the reagent includes one or more primer sets, wherein each primer set is configured to amplify nucleic acid sequences of bacterial species belonging to a set of bacterial species.

[0238] In some embodiments, one or more primer sets are selected from Table 1.1.

[0239] In some embodiments, amplification is PCR.

[0240] In some embodiments, the PCR is ddPCR, and the reagent further comprises one or more probes selected from Table 1.1.

[0241] In some embodiments, the bacterial species set includes one or more combinations of these bacterial species, wherein the combinations are selected from the group consisting of: Table 1.6.

[0242] In some embodiments, the bacterial species set consists essentially of: Actinomycete species oral taxa 181, Lawsonia non-saccharitroph, and Eubacterium species CAG: 274.

[0243] In some embodiments, the bacterial species set includes *Ruminococcus ovalis* (Broutella ovalis), *Lawsonia ovalis* (unsaccharitroph), *Roseidobacterium inulinum*, and optionally, *Actinomyces* species of oral taxa 181 and / or *Escherichia coli*.

[0244] In some embodiments, the bacterial species set includes Actinomycetes species oral taxa 181, Bacteroides fragilis, Escherichia coli, Lawsonia non-saccharitroph, Eubacterium species CAG: 274, Rosetteia inulinica, Rosetteia ovalis, Ruminococcus ovalis, and Dorgellella formica.

[0245] In some embodiments, the bacterial species set comprises the following: Actinomycete species oral taxa 181, Bacteroides fragilis, Escherichia coli, Lawsonia non-saccharitroph, Eubacterium species CAG: 274, Rosette of inulin, Rosette of enterica, Ruminococcus ovalis, and Dorperium formic acid-producing.

[0246] In some embodiments, a computer program product is provided for determining the risk of or diagnosing Crohn's disease (CD) in an individual, wherein the computer program product includes a computer-readable medium encoded with computer-executable code, wherein the computer-executable code is configured to perform the following steps: a) determining the relative abundance of a set of bacterial species in a fecal sample from the individual, wherein the set of bacterial species comprises one or more bacterial species selected from the group consisting of: *Dolphobic formate*, *Eubacterium pickyii*, *Roseinia inulinii*, *Roseinia enterica*, *Ruminococcus ovalis* (*Broutella ovalis*), *Eubacterium truncatula*, *Anaerobic bacillus hardruciformis*, *Lawsonia suis*, *Oscillatoria* species CAG: 241, *Oscillatoria* species 57-20, *Lactobacillus rhamnosus*, *Eubacterium* species CAG: 274, *Romebuts ileum*, *Bacteroides fragilis*, *Escherichia coli*, *Eubacterium gingivalis*, *Actinomyces* species oral taxonomy group 181, and *Actinomyces* species S6-Spd3; b) The relative abundance of each bacterial species in the bacterial species set from the individual is compared with the relative abundance of each bacterial species in the bacterial species set in the reference dataset to generate a risk score; and c) if the risk score is higher than a cutoff value, the individual is determined to have an increased risk of CD or to have CD.

[0247] In some embodiments, the reference dataset is obtained by determining the relative abundance of each bacterial species in a set of bacterial species from a subject group that includes CD patients and normal individuals in a reference cohort.

[0248] In some embodiments, a machine learning model is used to perform the comparison step in step (b) to generate a risk score.

[0249] In some embodiments, the machine learning model is constructed by: i) obtaining a reference dataset by determining the relative abundance of each bacterial species in a bacterial species set from fecal samples of a subject group comprising CD patients and normal individuals in a reference cohort; ii) dividing the reference dataset into a training dataset and a test dataset; iii) dividing the training dataset into a training subset and a validation subset; iv) training a candidate model using the training subset and evaluating the performance of the candidate model using the validation subset; v) optionally tuning the candidate model by adjusting the hyperparameters in the candidate model; vi) repeating steps (iv) and optionally (v) to generate a set of candidate models and selecting the best candidate model from the set of candidate models; vii) evaluating the performance of the best candidate model using the test dataset; viii) optionally training a final model with the reference dataset using a set of hyperparameters corresponding to the best candidate model to generate a final model; and ix) using the best candidate model from step (vii), or if the final model was trained using step (viii), using the final model from step (viii) as the model for step (b). The machine learning model in the comparison step is used to generate the risk score.

[0250] In some embodiments, the machine learning model is a random forest model, and the risk score in step (b) is generated by the following steps: (b1) generating a set of decision trees using the relative abundance of the bacterial species set in the reference dataset via the random forest model; and (b2) running the relative abundance of the bacterial species set from individuals along the set of decision trees to generate a risk score.

[0251] In some embodiments, the best candidate model is the candidate model with the highest performance in the candidate model set.

[0252] In some embodiments, the performance of a candidate model is evaluated by the area under the curve (AUC) of the receiver operating characteristic (ROC) curve of the candidate model.

[0253] In some embodiments, the cutoff value is in the range of 0.4 to 0.7.

[0254] In some embodiments, the cutoff value is 0.5.

[0255] In some embodiments, the cutoff value is determined by the following steps: (c1) calculating the Yoden index using the risk scores in the reference dataset; and (c2) determining the cutoff value based on the Yoden index.

[0256] In some embodiments, a method is provided for determining, diagnosing, preventing, or treating Crohn's disease (CD) in an individual, the method comprising the steps of: a) determining the relative abundance of a set of bacterial species in a fecal sample from the individual, wherein the set of bacterial species includes one or more bacterial species selected from the group consisting of: *Gimidae* formic acid bacteria, *Eubacterium hominis*, *Broutella ovalis*, *Roseobacterium inulinum*, *Clostridium sacchariformis*, *Eubacterium rectum*, *Roseobacter faecalis*, *Bacillus sacchariformis*, *Collinsella asiatica*, *Flavobacterium previae*, *Anaerobic bacillus*, *Trichophyton glomerulosa*, *Ruminococcus tortifolius*, *Clostridium truncatula*, *Bacteroides faecalis*, *Ruminococcus bryozoans*, *Roseobacterium intestinale*, *Adlerella cloacae*, *Alternaria putrefactive*, and *Eubacterium* species CAG. 38. *Rhodotorula hygroscopica*, *Butyrate-producing Agastobacterium*, *Formicate-producing Dorl*, *Bacteroides faecium*, *Vibrio vulnificus*, *Oscillatoria* species 57_20, *Cercozoella fecalith*, *Bacteroides vulnificus*, *Dolberella longiformis*, *Alternaria sarcoptera*, *Lawsonia sucrase*, Firmicutes bacteria CAG: 83, *Oscillatoria* species CAG: 241, *Eubacterium filariae*, *Alternaria finniformis*, *Bacteroides faecium*, *Ackermania myxophilus*, *Associated cocci*, *Vibrio virulentis*, *Alternaria inconspicuous*, *Bacteroides monomorpha*, *Bacteroides faecium*, *Eubacterium* species CAG: 274, *Desulfovibrio suis*, *Bacteroides masei*, *Clostridium* species CAG: 299, *Romebuts ileum*, *Romebuts ileum*, *Romebuts ileum*, *Anaerobic bacillus hardruciformis*, *Lactobacillus rhussi*, *Ruminococcus ovalis*, *Synthia scabra*, *Streptococcus lightensis*, *Actinomyces cariogenicus*, *Actinomyces* species S6-Spd3, *Actinomyces griseus*, *Lactobacillus mucosae*, *Streptococcus parahaemolyticus*, *Romenothiazus mucosae*, *Actinomyces* species oral taxa 181, *Veillonella septum*, *Escherichia coli*, *Bacteroides fragilis*, and *Romenothiza gingivalis*; b) compare the relative abundance of each bacterial species in the bacterial species set from the individual with the relative abundance of each bacterial species in the bacterial species set in the reference dataset to generate a risk score; c) if the risk score is higher than a cutoff value, determine that the individual has an increased risk of having CD or has CD; and d) if the individual is determined to have an increased risk of having CD or has CD, optionally treat the individual.

[0257] In some embodiments, the bacterial species set includes *Gimidae* formic acid bacteria, *Eubacterium hominis*, *Broutella ovalis*, *Roseobacterium inulinum*, *Clostridium sacchariformis*, *Eubacterium rectum*, *Roseobacterium falciparum*, *Bacillus crypticarum*, *Collinsella aerogenes*, *Bacillus preibryophyte*, *Anaerobic bacillus hardruciformis*, *Trichophyton molasses*, *Ruminococcus tortifolius*, *Clostridium truncatella*, *Bacteroides fecalis*, *Ruminococcus bryophyte*, *Roseobacterium entericarum*, *Adlerococcus brevis*, *Alternaria putrefactive*, *Eubacterium* species CAG: 38, *Roseobacter humanis*, *Agastache rugosa*, *Dolphobicella formate*, *Bacteroides fecalis*, *Ostomyces viscera*, *Vibrio spirochetes* species 57-20, *Cercospora fecalis*, *Bacteroides var. ... 241. *Eubacterium filamentosa*, *Alternaria fenestrate*, *Bacteroides foetida*, *Ackermania spp.*, *Concomitant cocci*, *Toxic butyrosporum*, *Alternaria inconspicuous*, *Bacteroides monomorpha*, *Bacteroides foetida*, *Eubacterium* species CAG: 274. *Desulfovibrio suis*, *Bacteroides masei*, *Clostridium* species CAG: 299. *Hemolytic twin cocci*, *Streptococcus pyogenes*, *Actinomyces cariogenicus*, *Actinomyces* species S6-Spd3, *Actinomyces griseus*, *Lactobacillus mucosae*, *Streptococcus parahaemolyticus*, *Rhodopseudomonas mucosae*, *Actinomyces* species oral taxonomy group 181, *Veillonella spp.*, *Escherichia coli*, and *Bacteroides fragilis*.

[0258] In some embodiments, a method for monitoring Crohn's disease (CD) activity in an individual is provided, the method comprising the steps of: (a) determining the relative abundance of a set of bacterial species in a fecal sample from the individual, wherein the set of bacterial species comprises one or more bacterial species selected from the group consisting of: formic acid-producing Dorperella, ovale-forming rumenococci (ovale-forming Broutella), sugar-insoluble Lawsonia, and Escherichia coli; (b) comparing the relative abundance of each bacterial species in the set of bacterial species from the individual with the relative abundance of each bacterial species in the set of bacterial species in a reference dataset to generate a risk score; and (c) if the risk score is higher than a cutoff value, determining that the individual has increased disease activity.

[0259] In some embodiments, the reference dataset is obtained by determining the relative abundance of each bacterial species in a set of bacterial species from a subject group that includes active CD patients and inactive CD patients in a reference cohort.

[0260] This invention provides a set of bacterial biomarkers (used as individual biomarkers or in combination as shown in Tables 1.4, 5 and 9) for determining the presence of CD in subjects or assessing the risk of CD. The set of microbial biomarkers may include (1) all microbial biomarkers listed in Tables 1.1-1.5, and (2) one or more microbial biomarkers selected from Tables 1.1-1.5.

[0261] Risk assessment or diagnosis of Crohn's disease

[0262] 1. Workflow

[0263] In one embodiment, to determine the risk of Crohn's disease in an individual or whether the individual has Crohn's disease, the following steps are performed: (1) obtaining a fecal sample from the individual and determining the level / relative abundance of one or more bacterial species selected from Tables 1.2 and 1.3 (e.g., in combinations specified in Tables 1.4 or 1.5). (2) obtaining a fecal sample from a reference cohort containing subjects with or without Crohn's disease and determining the level / relative abundance of the same bacterial species as in (1). (3) comparing the level / relative abundance of the bacterial species obtained in (1) with the level / relative abundance of the bacterial species obtained in (2).

[0264] In some embodiments, a set of bacterial biomarkers (used as individual biomarkers or in combination as shown in Tables 1.4, 5 and 6) is provided for use in determining the presence of CD in subjects or assessing the risk of CD. The set of microbial biomarkers may include (1) all microbial biomarkers listed in Tables 1.1-1.5, and (2) one or more microbial biomarkers selected from Tables 1.1-1.5.

[0265] 2. Determine the levels of bacterial markers.

[0266] In one embodiment, a method is provided that measures the level or amount of characteristic DNA or RNA of one or more bacterial species found in a human fecal sample as a means of diagnosing or assessing Crohn's disease risk. Therefore, the first step of the method is to obtain a fecal sample from a test subject and extract microbial DNA or RNA from the sample.

[0267] Fecal sample acquisition and preparation

[0268] The exemplary methods disclosed herein are used to obtain stool samples from individuals undergoing testing or monitoring for Crohn's disease. Stool samples from individuals can be easily collected, whether in a clinic or at the patient's home. An appropriate amount of stool is collected and stored according to standard procedures for further preparation. According to the exemplary methods, established techniques can be used to analyze bacterial DNA or RNA found in the patient's stool samples. Methods for preparing stool samples for nucleic acid extraction are well known to those skilled in the art.

[0269] DNA and RNA extraction and quantification

[0270] The methods used to extract DNA from biological samples are well-known and standard practice in the field of molecular biology. RNA contamination should be eliminated to avoid interfering with DNA analysis.

[0271] Similarly, there are many methods for extracting DNA from biological samples. The most common DNA extraction methods are mechanical, chemical, and enzymatic lysis, precipitation, purification, and concentration. Other specific methods for DNA extraction include phenol-chloroform extraction, alcohol precipitation, or silica-based purification. Commercial kits, such as the MO BIO PowerSoil DNA Isolation Kit (MO BIO Laboratories, Carlsbad, CA, USA), the QIAamp DNA StoolMini Kit (Qiagen, Hilden, Germany), and the Maxwell® RSC PureFood GMO and Identification Kit (Promega), can also be used to obtain DNA from biological samples of test subjects. The quality and concentration of the extracted DNA will be further evaluated.

[0272] PCR-based quantitative determination of bacteria

[0273] Once DNA is extracted from a sample, the amount of predetermined bacterial DNA (e.g., 16S rDNA or DNA encoded by bacterial genes specific to the bacterial species) is quantified. In some embodiments, the method for determining DNA levels is based on amplification methods, such as polymerase chain reaction (PCR) such as real-time quantitative PCR or droplet digital PCR (ddPCR), for DNA quantification analysis.

[0274] General methods for PCR are well known in the art and will not be described in detail here. PCR reagents and protocols are also available from commercial suppliers such as Roche Molecular Systems.

[0275] PCR is most commonly performed as an automated process using thermostable enzymes. In this process, the temperature of the reaction mixture is automatically cyclicated through the denaturation zone, primer annealing zone, and extension reaction zone. Machines specifically modified for this purpose are commercially available.

[0276] Real-time quantitative PCR (a semi-quantitative PCR method) measures fluorescence after each cycle, and the intensity of the fluorescence signal reflects the instantaneous amount of DNA amplicon in the sample at that specific time.

[0277] Droplet-based digital PCR (ddPCR) is an improvement on the conventional polymerase chain reaction (PCR) method. In ddPCR, DNA / RNA is randomly encapsulated within droplets that serve as the reaction chamber. It provides more sensitive nucleic acid detection and absolute quantification.

[0278] Although PCR amplification of target bacterial DNA or RNA has been used in some of the embodiments disclosed herein, those skilled in the art will recognize that amplification of these DNA or RNA species in a sample can be accomplished by any known method, such as ligase chain reaction (LCR), transcription-mediated amplification and self-sustaining sequence replication or nucleic acid sequence-based amplification (NASBA), each of which provides sufficient amplification. Recently developed branched DNA techniques can also be used to quantify the amount of DNA or mRNA in a sample.

[0279] Bacterial Quantitative Determination Based on Metagenomics

[0280] Metagenomic datasets can include sequencing data generated from the same standardized protocol, encompassing steps from fecal DNA extraction to sequencing, raw data quality filtering, host read decontamination, and microbiome interpretation. Various techniques, such as shotgun metagenomic sequencing, can be used to characterize the species in a sample. In shotgun metagenomic sequencing, DNA is obtained from heterogeneous cellular samples and fragmented into DNA fragments that can be compared with the genomes of multiple microbial species to identify the species in the sample.

[0281] 3. Generation of Disease Risk Score (Disease Probability)

[0282] Machine learning models

[0283] In one embodiment, a machine learning model, such as a random forest model, can be used to generate disease risk scores (disease probabilities). Random forest is a supervised learning approach for machine learning, used for classification and regression. It is a supervised machine learning algorithm that improves the prediction accuracy of a dataset by averaging or making a final decision based on the majority votes of many decision trees applied to different subsets of the dataset.

[0284] In this regard, in one embodiment, the method includes constructing a machine learning model by: (1) obtaining a reference dataset in a reference queue by determining the relative abundance of bacterial species selected from Tables 1.2 and 1.3 (e.g., combinations specified in Tables 1.4 or 1.5) in a fecal sample; (2) splitting the data into a training dataset and a test set; (3) splitting the data into a training subset and a validation subset under the training dataset, training a candidate model using the training subset, and evaluating the performance of the candidate model using the validation subset; (4) tuning the candidate model by using different combinations of hyperparameters; (5) repeating steps (3)-(4) and selecting the best candidate model with the best performance; (6) evaluating the candidate model using the test set; and (7) as an optional step, training a final model using the entire reference dataset and the hyperparameter combination of the best candidate model to obtain the final model; (8) using the best candidate model from step (6), or if step (7) is used to train the final model, using the final model from step (7) as the machine learning model.

[0285] After constructing the machine learning model, the model can be deployed to generate a risk score (disease probability) for a subject whose Crohn's disease risk is to be determined. In one embodiment, the method includes: (1) determining the relative abundance of the corresponding bacterial species in the individual for whom a Crohn's disease diagnosis is to be determined (the same as that used to construct the machine learning model); (2) inputting the relative abundance of these species from the individual obtained in step (1) into the constructed machine learning model to generate a risk score (disease probability); if a random forest model is used, the relative abundance of the species listed in Tables 1.2 and 1.3 obtained from the individual in step (1) is run along a decision tree in the random forest model; (3) when the risk score is greater than a cutoff value, determining that the individual has a risk of Crohn's disease or has Crohn's disease, and when the risk score is less than a cutoff value, determining that the individual has no risk of Crohn's disease or does not have Crohn's disease.

[0286] In some embodiments, the cutoff value can be determined as follows.

[0287] Determination of cutoff value

[0288] After obtaining the subject's disease risk score (disease probability), a default cutoff value of 0.5 or an optimized cutoff value can be used to determine whether the subject has Crohn's disease or is at risk of developing Crohn's disease.

[0289] In some embodiments, an optimized cutoff value can be calculated based on the Youden index using risk score data from a reference cohort. In some embodiments, the risk score data is obtained by generating a risk score for each bacterial species by inputting the relative abundance of bacterial species in the test set of the reference cohort into a machine learning model. The Youden index can integrate sensitivity and specificity information, and by using Youden index analysis, an optimal cutoff value that provides the best trade-off between sensitivity and specificity can be obtained. The Youden index (Y) can be calculated using the following formula:

[0290] Example methods for determining CD risk in subjects

[0291] In one embodiment, to determine a Crohn's disease risk score in a subject, the following steps are performed: (1) measuring the relative abundance of a selected set of bacterial species in the subject; (2) inputting the relative abundance data into a pre-trained random forest model to generate a disease risk score (disease probability) for the subject; and (3) if the subject's risk score (disease probability) is higher than an optimized cutoff value, the subject is considered to have Crohn's disease or be at risk of having Crohn's disease. Conversely, if the risk score (disease probability) is not higher than an optimized cutoff value, the individual is considered not to have Crohn's disease or be at risk of having Crohn's disease. Example 1: Methods (sample collection, DNA extraction, and sequencing) Cohort description and study subjects

[0292] In this study, shotgun metagenomic profiling of fecal microbiomes from three different cohorts was analyzed to discover and validate diagnostic models constructed using Crohn's disease (CD)-specific bacterial species biomarkers. In the discovery cohort, 292 Chinese participants (aged 18–87 years) were recruited, including 174 patients with Crohn's disease (CD) and 118 healthy individuals. Males comprised 53.1% of the total participants. In the validation cohort, 200 Chinese participants (aged 20–73 years) were recruited, including 92 patients with CD and 108 healthy individuals. Males comprised 59.6% of the total participants. In another validation cohort, 119 Australian participants (aged 18–78 years) were recruited, including 98 patients with CD and 81 healthy individuals. Males comprised 46.6% of the total participants.

[0293] Patients were included if they were 18 years of age or older and had a diagnosis of CD as defined by endoscopy, radiology, and histology; and were on stable medication. Patients were excluded if they had: an infection with an intestinal pathogen; short bowel syndrome; or severe liver, kidney, endocrine, respiratory, neurological, or cardiovascular disease.

[0294] Healthy individuals without serious diseases (such as inflammatory bowel disease, cancer, or advanced adenoma) were included as controls.

[0295] All participants consented to donate stool samples and participate in a questionnaire survey, for which written informed consent was obtained. Stool samples from study participants were stored at -80°C for downstream microbiome analysis.

[0296] In addition, three public datasets containing data from a total of 248 participants were downloaded for model validation. These participants included 102 participants from the United States (68 CD patients and 34 healthy individuals, aged 21 to 82 years), 44 participants from the Netherlands (20 CD patients and 22 healthy controls, aged 21 to 71 years), and 102 participants from China (48 CD patients and 54 healthy controls, aged 13 to 51 years). Fecal DNA extraction and DNA sequencing

[0297] Fecal bacterial DNA was extracted using a modified standard protocol (Maxwell® RSC PureFood GMO and Identification Kit, Promega) to improve DNA yield. Approximately 100 mg of each fecal sample was pretreated: the sample was suspended in 1 ml ddH2O and precipitated by centrifugation at 13,000 × g for 1 min. The washed sample was thoroughly mixed with 800 μL TE buffer (pH 7.5), 16 μL β-mercaptoethanol, and 250 U lysozyme, and digested at 37°C for 90 min, followed by centrifugation at 13,000 × g for 3 min to precipitate the remaining DNA.

[0298] After pretreatment, the precipitate was resuspended in 800 μL of CTAB buffer (Maxwell® RSC PureFoodGMO and identification kit, according to the manufacturer's instructions) and thoroughly mixed. The samples were heated at 95°C for 5 minutes and then cooled, followed by vortexing at 2850 rpm for 15 minutes using 0.5 mm and 0.1 mm beads to release nucleic acids from the samples. Subsequently, 40 μL of proteinase K and 20 μL of ribonuclease A were added, and the nucleic acids were digested at 70°C for 10 minutes. Finally, the supernatant was obtained after centrifugation at 13,000 × g for 5 minutes and transferred to a Maxwell® RSC instrument for DNA extraction. The extracted fecal DNA was used for ultra-deep metagenomic sequencing via an Ilumina Novaseq 6000 for in-house (MagIC, Hong Kong) or outsourced to a company (Novogene, Beijing, China). Quality control of raw sequences

[0299] The original sequence reads were pruned using Trimmomatic (v0.39). Non-human reads were then separated from contaminated host reads. The steps to obtain clean reads included: 1) removing adaptors; 2) scanning the reads using a 4-base-wide sliding window and removing reads when the average quality per base dropped below 20; and 3) reducing the read length to below 50 bases. The pruned sequence reads were mapped to the human genome (reference database: hg37decv0.1) using KneadData (v0.10.0) to remove host-derived reads. The two reads at their paired ends were tandemly joined. Example 2: Methods for determining the composition of the human gut microbiota and the differentially expressed bacterial species between CD patients and normal individuals Bacterial microbiome analysis

[0300] Bacterial community composition profiles were analyzed using metagenomically pruned reads via MetaPhlAn3 (v3.0.13). Reads were mapped to clade-specific biomarker genes and species pangenomes were annotated using Bowtie2 (v2.2.4.2). The output table contains bacterial species and their relative abundance at different levels (from kingdom to strain). The data were analyzed in R v3.6.1 using ggpubr (v0.2) and phyloseq (v1.24.2). Human gut bacterial composition and selected differentially expressed bacterial species were compared between CD patients and healthy individuals using multivariate linear model association analysis (MaAsLin2). Machine learning models

[0301] MaAsLin2 is used for the identification of discriminative features. Random Forest (RF) is chosen to build a predictive model for CD patients compared to normal controls using selected fecal microbiota, as RF has superior performance in binary feature classification. Random Forest is one of the approaches to building predictive models in metagenomic data analysis. As a widely used ensemble learning algorithm, Random Forest consists of a series of Classification and Regression Trees (CART) to form a powerful classifier. A subset of data randomly sampled from the original dataset with replacement is called bootstrap sampling, which is suitable for building trees. When the training dataset for the current tree is plotted using the bootstrap method, Observations are excluded from the overall dataset. In the case of infinite N, 36.8% of the data does not appear in the training samples, known as out-of-bag (OOB) observations, which will not be used to build the tree. Furthermore, additional randomness is introduced into the random forest because each decision tree splits nodes based on a random subset of features selected from the overall features. In each iteration, nodes are split to generate a tree using features with the smallest Gini coefficient (the Gini coefficient is used to evaluate the purity of a node). Using different subsets of data and features, the algorithm is able to train different trees and obtain the final classification by averaging or majority voting the results from the tree models.

[0302] A total of 174 CD patients and 118 healthy controls were included as the discovery cohort for modeling. Selected species identified by MaAsLin2 were used as input for random forest model construction. Model performance was evaluated using 5-fold cross-validation. For binary classifiers, the area under the receiver operating characteristic (ROC) curve (AUC) was used to evaluate model performance. The parameters of the model construction were tuned to achieve optimal accuracy and κ. These analyses were performed using the R packages MaAsLin2 v2_1.4.0, randomForest v4.6-14, and pROC v1.18.2. Droplet digital PCR (ddPCR) Primer and probe design for nine bacterial species

[0303] Specific gene sequences for each bacterium were downloaded from the MetaPhlAn3 database, and the specificity of each bacterial gene sequence was confirmed by performing a BLAST procedure using a public database such as GenBank. Based on the gene sequences, primers and probes were designed in Primer3Plus (https: / / www.primer3plus.com / index.html) by evaluating the Tm value, GC content, and possible secondary structure of the primers and probes. The primer and probe sequences used for the 16S rDNA internal control were identical to those reported in previous studies. Primers and probes were synthesized by BGIBio-Solutions Hong Kong Co., Limited. Detection of selected bacterial species using droplet digital PCR (ddPCR).

[0304] Microdroplet digital PCR reactions were designed to detect selected bacterial species (listed in Tables 1.2 and 1.3). The first reaction was designed to detect *Bacteroides fragilis* (probe concentration 120 nM), *Eubacterium* species CAG:274 (probe concentration 400 nM), *Escherichia coli* (probe concentration 120 nM), and *Actinomyces* species oral taxa 181 (probe concentration 400 nM). The second reaction was designed to detect *Roseidone* (probe concentration 170 nM), *Ruminococcus ovalis* (probe concentration 480 nM), *Roseidone* enterica (probe concentration 180 nM), and *Lawsonia lactiflora* (probe concentration 450 nM). The third reaction was designed to detect *Dolphinium formate* (probe concentration 250 nM) and the 16S reference gene (probe concentration 350 nM). Primer and probe sequences used to detect certain identified bacterial species (including the nine selected bacterial species described herein) and the reference gene are listed in Table 1.1. The ddPCR mixture consisted of the following: 10 μL of ddPCR probe premix (Supermix for Probes) (dUTP-free, Bio-rad, catalog number 1863024), primers (900 nM), probes (concentrations listed herein), 2 μL of DNA (diluted to 5 ng / μL in reactions 1 and 2, and to 0.05 ng / μL in reaction 3), and nuclease-free water (to 20 μL). Then, 20 μL of the ddPCR mixture and 70 μL of probes were loaded into a DG8™ box (Bio-rad, catalog number 1864008) using Droplet Generation Oil for Probes (Bio-rad, catalog number 1863005). A DG8™ gasket (Bio-rad, catalog number 1863009) was hooked onto the box holder. Droplets for each sample were generated using a QX200 droplet generator (Bio-Rad Laboratories) and then transferred to ddPCR™ 96-well plates (Bio-Rad Laboratories catalog number 12001925). After sealing the plates with aluminum foil (Bio-Rad Laboratories catalog number 1814040) and PX1 PCR plate sealant (Bio-Rad Laboratories), the 96-well plates were run on a Bio-Rad Laboratories T100 PCR system. The PCR program was as follows: (1) pre-denaturation at 95°C for 10 min; (2) denaturation at 94°C for 30 s, followed by annealing and extension at 59°C for 1 min, for 40 cycles; (3) enzyme inactivation at 98°C for 10 min. After PCR, the fluorescence of the droplets was detected in a QX200 droplet reader (Bio-Rad Laboratories), and the data were analyzed using QuantaSoft™ Analysis Pro (v 1.0.596). The concentration of each bacterial species was determined and then normalized using the concentration of the 16S reference gene (normalized abundance of the target = target concentration / 16S concentration). Table 1.1: Nucleotide sequences of primers and probes for bacterial species Table 1.1 lists the nucleic acid sequences of the PCR products amplified by the primers targeting the selected bacterial species.

[0305] Table 1.1 lists the nucleic acid sequences of the PCR products (“amplified fragments”) amplified by the primers for the selected bacterial species, as well as the nucleic acid sequences of the genes containing these amplified fragments (“marker genes”).

[0306] Bacteroides fragilis

[0307] Amplified fragment: Region 1564367 to 1564483 of Bacteroides fragilis NCTC 9343, complete genome (GenBank: CR626927.1)

[0308] Nucleic acid sequence:

[0309] TCTGGTTATCCGGACTCCTTCCGGGCATCCGATGG CGGACGTGGACATACCGCTTCGCTTTGCCGCTATGGTGCCGTTGCAGGTTTACAACACACTGCAGGTACCGGCAGAAAAAGA (SEQ ID NO: 58)

[0310] Biomarker gene: menE (1095 nt)

[0311] Gene product: putative O-succinylbenzoate-CoA ligase

[0312] Nucleic acid sequence:

[0313]

[0314] E. coli

[0315] Amplified fragment: Region 2558008 to 2558115 of chromosome 25922 of Escherichia coli strain ATCC, complete genome (GenBank: CP117235.1)

[0316] Nucleic acid sequence:

[0317] AATGACTGGTGCATTCCGAAGGGAATAACGCCA ACGACGCACATTATCAAATTGCCGATTGGCGAAATCAGACAGCCCAATGCGACGCTCGATCTTAGCCAAAGCGTT (SEQ ID NO: 59)

[0318] Biomarker gene: hipA (1323 nt)

[0319] Gene product: Type II toxin-antitoxin system serine / threonine protein kinase toxin HipA

[0320] Nucleic acid sequence:

[0321]

[0322] Dauerella formitogenes

[0323] Amplified fragment: Region 1799237 to 1799333 of chromosome 27755 of formic acid-producing strain Dornos formic acid-producing bacteria, complete genome (GenBank: CP102279.1).

[0324] Nucleic acid sequence:

[0325] CAGAGAATGGCAGACGGAAATCTTGCAGACTTGA TCCTGACAACAGGCGGAGCAGGCTGCGGACCGGAAGACTGCACACCGGAGGCAACTATGGAAG (SEQ ID NO: 60)

[0326] Biomarker gene: NQ560_08855 (471 nt)

[0327] Gene product: MogA / MoaB family molybdenum cofactor biosynthesis protein

[0328] Nucleic acid sequence:

[0329] TTATTTTATTCCCTGTGCCACTTCAACTGCATGCA TGACCTCTGGCAAGATGTATCCAAGACATTCCTTTACAGCTCCTGCTTTTCCAGGAAGATTTACAATCACAACATTTCCACGGATTCCTGCCGCTGCACGGTTCAACATAGCTCTTTTTGTCAGTGTAAGCATGTGTGCACGCATACCTTCCGGAATTCCAAGAAGTGGTCTGTCTACAACTTCCATAGTTGCCTCCGGTGTGCAGTCTTCCGGTCCGCA GCCTGCTCCGCCTGTTGTCAGGATCAGGTCTGCAAGATTTCCGTCTGCCATTCTCTGAAAAACAGTAGATAATACCGTTTTGTCTGTCGGGAGAGCCTTCATGAACACAACGTCAAGCCCTGCTTCCTCCACCATTCTGCGGATTACTTTTCCACTTGCATCCTCCTGCTCTCCTTTATAAACTGCTGTACTTGCAGTAATAATCGCTACTTTCAT (SEQ ID NO: 78)

[0330] Picky Eubacterium (Picky Helichrysum)

[0331] Amplified fragment: Regions 1247080 to 1247192 of chromosome 1570 of the *Trichophyton spp.* strain FDARGOS_1570, complete genome (GenBank: CP085938.1).

[0332] Nucleic acid sequence:

[0333] GCCAAAGGTTGCACATCCATATGTGCAGGAGTATG TGGCTGGCAGAGAGCATGGCAATGTAAAAGTACAGGATTTTCTTGATCATGGAGAAGAATTTGCAATTGGGGCAGATG (SEQ ID NO: 61)

[0334] Biomarker gene: LK414_08735 (429 nt)

[0335] Nucleic acid sequence:

[0336] ATGAAAAAAGTATTTTTAATTGGGTTTACATTGAT GTGTTTTGCTTCTATGGCAGGTTGTAAGACAAAGCCGCCAAAGGTTGCACATCCATATGTGCAGGAGTATGTGGCTGGCAGAGAGCATGGCAATGTAAAAGTACAGGATTTTCTTGATCATGGAGAAGAATTTGCAATTGGGGCAGATGAAAATGGAAAAGCGGTATTTAAAGACCCGCAGAAAGCATTTGAGGCGTTA GTAAGGGATTATTCAGATGGTATTAATCTGATTAGGGATGAGTACAAGCTTGGACCAATCACTCCAAGAAACTTTTATGATTACATGACGTATGGGTATCAGGTTAATACTGGCACAGAAGAAAGCAAGAATCAGGCTGCTTTTGTGACTCAGGTTTTAGATATATATGAGAATAGTTATGATTTTGATAAATAG (SEQ ID NO: 79)

[0337] Inulin-producing bacteria

[0338] Amplified fragment: Whole genome shotgun sequence of region 1160270 to 1160376 of *Rossiella pneumoniae* DSM 16841 strain FDARGOS_1587 ctg.s1.000001F_arrow_pilon (GenBank: JAJFOJ010000002.1)

[0339] Nucleic acid sequence:

[0340] GGAACCAATGGAAAACTGTGTCAGGGCAGTTCCG GACAAACAGCCGGTCATTGTTCCGGGAAGCGTGTTTGGCAGGGACGGCAGCCGTTATGGATATGGAAAGGGTT (SEQ ID NO: 62)

[0341] Biomarker gene: LK408_RS18325 (567 nt)

[0342] Gene product: 5-formyltetrahydrofolate cyclase

[0343] Nucleic acid sequence:

[0344] CTATACTCCCTTTCTGATCACTTCTGTCTCGGTATAT (SEQ ID NO: 80)

[0345] Enteric Rosebene bacteria

[0346] Amplified fragment: Regions 1246072 to 1246169 of the genome assembly of Enterobacter rosenbergii L1-82, chromosome: 1 (GenBank: LR027880.1)

[0347] Nucleic acid sequence:

[0348] TCGGATGGATCGTCTATCTGATCTGCCTGCAGTTT GACACCGGAACCGTCATTGCAAACATGATCGCAACACTTGCCCTCACGGTCTTTTCAAGGATG (SEQ ID NO: 63)

[0349] Biomarker gene: RIL182_01152 (456 nt)

[0350] Nucleic acid sequence:

[0351] TCAGGCATGGCTTCTTTGCCGTTTTTGTCCGGTGCT GCAGTGCGTTAAACCATGCCTGCGGCAGGGAAAAACCGAAGATGATACCGAGTACGATCGCTCCGGCAACTTTGATTGTCTCCATTCCGTAACGTGAAAATTCATTCATGTCATTCATAATAAAATAATAGGAAGTATAATAAATACCAGCCCCGGCACCAGTGGGAAAATACCGGCGATCAGGTAGACGGTCACCGGATTTTTACGCAGG GCGGCAACCATCCTTGAAAAGACCGTGAGGGCAAGTGTTGCGATCATGTTTGCAATGACGGTTCCGGTGTCAAACTGCAGGCAGATCAGATAGACGATCCATCCGATCGCACCGGTAAGTCCGCAGAAGAGAAGTTCTGATTTTGGTGCACAAAACAGCACGGCAAAAGAAAGCGTTGCGATCATACTGACAATAAACTGAATAAACAT (SEQ ID NO: 81)

[0352] Oval rumenococci (Oval Broutella)

[0353] Amplified fragment: Regions 3464821 to 3464916 of chromosome 29174 of *Brutella ovalis*, complete genome (GenBank: CP102265.1).

[0354] Nucleic acid sequence:

[0355] CACCTGGATACCGTAAAGAACGGTGGAAAATATG ATGGCGCACTGGGCGTTGTGACTGGTGTGGCAGCGCTGGGATACCTCAAACAGTCGGGCTTC (SEQ ID NO: 64)

[0356] Biomarker gene: NQ503_16555 (1236 nt)

[0357] Gene product: M20 family metallohydrolases

[0358]

[0359] Protruding belly eubacterium

[0360] Amplified fragment: Region 1481183 to 1481301 of chromosome 27560 of *Eubacterium truncatum* strain ATCC, complete genome (GenBank: CP102282.1).

[0361] Nucleic acid sequence:

[0362] GGACAGCCGGCAAAAATATTTCAGGGAATCATGG GACTATGTCAGGGTATATTCGTAGGTTGCCTTGCAGTGGCACTGGCGGAGGCCCTAAACGGAACAGCAATCTTTGCACGAAGGGC (SEQ ID NO: 65)

[0363] Biomarker gene: NQ558_06695 (432 nt)

[0364] Gene product: V-stage sporogenin AB

[0365] Nucleic acid sequence:

[0366] ATGATAGGAAAAATAATTCTTGCAATTACAGGAA TCGGGGCAGGTTTCATAATCTCTGCAGGAGTTTTTGCATTAATAACATCAACAGGAATAATAACAAGATTTGCAGACAAAACCCATACTGCAAAAAGTATACAATTATATGAAACAATAGTAATTGCAGGGGGAATACTTTGGAATATATTCTGGGTAATGGAATCCCATTTCAAATTCACGGGACAGCCGGCAAAAATAT TTCAGGGAATCATGGGACTATGTCAGGGTATATTCGTAGGTTGCCTTGCAGTGGCACTGGCGGAGGCCCTAAACGGAACAGCAATCTTTGCACGAAGGGCAAAACTGCAAATGGGGCTAAGTTTCATAGTGTTATCAGTAGCGGTAGGCAAAGTCCTTGCATCGTTGCTCCAATTTTACAACGACTGGGTGAACTAA (SEQ ID NO: 83)

[0367] Hadrulina anaerobic bacillus

[0368] Amplified fragment: Regions 2653461 to 2653547 of the DNA of *Hardrophyte JCM 17467*, complete genome (GenBank: AP028031.1)

[0369] Nucleic acid sequence:

[0370] AACAATCTGCCATCGTGTTTTCATGGTGGCGGG TATATTGGGGGCAGCACTTTGGTCTTACAAAATCAATGCCGTTTTCTTGCTG (SEQ ID NO: 66)

[0371] Biomarker gene: Ahadr17467_25260 (1041 nt)

[0372] Gene product: α / β hydrolase

[0373] Nucleic acid sequence:

[0374]

[0375] Lawsonia sucrase

[0376] Amplified fragment: Regions 627394 to 627504 of chromosome OA10 of *Lactobacillus suis* strain, complete genome (GenBank: CP091870.1).

[0377] Nucleic acid sequence:

[0378] TTGACGCCTTGACCATATCCACTTTGAAGCACCTA CTGCGCCAGCTTCCGCTGCGTACGGACATTACCATGGAGGAAGTGGTGGACCTATTCCGGCAGTTTCAGGACTTTA (SEQ ID NO: 67)

[0379] Biomarker gene: L9O85_02835 (633nt)

[0380] Gene products: TetR / AcrR family transcription regulators

[0381] Nucleic acid sequence:

[0382] TTAGTAGCCATATCTATCCCCCCTTTCGGTGACGC CGTAGAGCAGGACAGTCAAAAGCCGGCGGCACTGGGCATCCCGAAGAGCAAAGCACTTCCTGGCAGGCATGGAAGATGCTTGCGCCTGCGCATTGATAAAGTCCTGAAACTGCCGGAATAGGTCCACCACTTCCTCCATGGTAATGTCCGTACGCAGCGGAAGCTGGCGCAGTAGGTGCTTCAAAGTGGATATGGTCAAGGCGTCAAACGCCTTCCGACGTGCTTGGATCTGTTCGAACAGCCGGGCGGGAGGAGAGATGACTGCTTCGCAGAAGATGGGTTGATAGACCGGATTCTCCTGAAAAAAGGTCATGCGGGCAGTAAAATACCGCCCCAGCTGGTCCTCCACAGAGCCGTCGTCCTCCTCCGGCAGAGCAGCGCTGAGATGGGCGGTGAGGCGCTGAAAGCACTCCTCCACACAGGCCAGATATAATTCGTCCTTGTTCTCAAAATAGTGGTAGATGATCCCTTTGGAAATTCCTTGGGCGGCACAGATGGTGTTTACCGAACTGCCGCCGTGACCCTTTTCCGCAAATTCCGCCAGTGCGCTGTCCATAATCCTACGCTGGGTCTGTTGATTTTTTTCTTCCCGTTTCAT (SEQ ID NO: 85).

[0383] CAG: 241 of the genus *Vibrio*

[0384] Amplified fragment: 1167 to 1283 region of Oscillospiraceae species CAG:241, WGS project CBDI01000000 data, contig, whole genome shotgun sequence (GenBank: CBDI010000003.1)

[0385] Nucleic acid sequence:

[0386] ACAGCCTCTGGTCATCCTTATACTCCACGTAGGCAT TGGCGTCATCCGCGAAAAAGTAATTCGCGTAGTCTTTTTTGTAACTATTAGAAAATGCCACGGGGTATCTGATATTCGCGG (SEQ ID NO: 68)

[0387] Biomarker gene: BN557_01301 (1506 nt)

[0388] Nucleic acid sequence:

[0389]

[0390] Species of the genus *Vibrio* 57_20

[0391] Amplified fragment: region 33675 to 33760 of MAG: Oscillospiraceae species 57_20 Ley3_66761_scaffold_1931, whole genome shotgun sequence (GenBank: MNTE01000025.1)

[0392] Nucleic acid sequence:

[0393] TGAAAATGGCCCAATCCCTTATGGGAAACGGCGG TGGGACGGAGCAGCCTCCCCCCGCCGCGCCGCCCTCCGGCGAAACAGGAAAC (SEQ ID NO: 69)

[0394] Marker gene: BHW41_06875 (522 nt)

[0395] Nucleic acid sequence:

[0396] TTATACATGGCCATCCCCCCTCAGCAGGAACTTTTT CCCCACCCGGAACAGCCGGGCCAGCTGCAATGCCCGTTCCACCTTATCCTGCCGCTCCGGTTTCAGATAGGGCCGCAGTGCGAACAAAAGCTGACGGGCGTTGCTGTTCTGGGCCTCTCCCAACTCCCGGAGGACCGGCAGGAGGGAGGCGATGGCCTTGGGATCCAGATCACCCAGTCCGGAGAGAAAACCTGGGCCAGACCCCGGCTGACCGCTCGCCACAGGCCCGCTGGGAGACGGGATTCGGCTTTCCGGGGTGGGCTGGCCCCAACTCTGTGCCGTCTCCCTGCTGGCGGGACCGCTCTGCTGAAACGGAAGGCCGCCGTTCTGCGCTGCGTTTCCTGTTTCGCCGGAGGGCGGCGCGGCGGGGGGAGGCTGCTCCGTCCCACCGCCGTTTCCCATAAGGGATTGGGCCATTTTCATGATCTGGTCCATGGCCTCCGGGTTGGAGAGGAGACTATTCAGCTTGTCGTCAAACTCCGCCAT (SEQ ID NO: 87)

[0397] Lactobacillus rhubarb

[0398] Amplified fragment: Region 22 to 132 of the 16S ribosomal RNA of Lactobacillus roguesi strain ATCC 27753 (partial sequence NCBI reference sequence: NR_104836.1)

[0399] Nucleic acid sequence:

[0400] TGCTTAACACATGCAGTCGAACGAAGCATTTAAG ACAGATTACTTCGGTTTGAAGTCTTTTATGACTGAGTGGCGGACGGGTGAGTAACGCGTGGGTAACCTGCCTCATAC (SEQ ID NO: 70)

[0401] Biomarker gene: rRNA-16S ribosomal RNA (1452 nt)

[0402] Gene product: rRNA-16S ribosomal RNA

[0403] Nucleic acid sequence:

[0404]

[0405] Eubacterium species CAG: 274

[0406] Amplified fragment: Region 7396 to 7500 of Eubacterium species CAG:274, WGS project CBEX01000000 data, contigs, whole genome shotgun sequence (GenBank: CBEX010000022.1).

[0407] Nucleic acid sequence:

[0408] TCATTTGTTTGGGTGCTGTCTTAGGTCTTGTGTTGG TTACAAAAATAGCCAAGGCTGAGGGGCAGAACCCTGATACATATGTGGACTTCGCTATATGGGCTTTGA (SEQ ID NO: 71)

[0409] Biomarker gene: BN582_00070 (861 nt)

[0410] Gene product: prolipoprotein diacylglycerol transferase

[0411] Nucleic acid sequence:

[0412] TTATTCTCCGTTTTTCTTATTTTTCACAAGCATTAC(SEQ ID NO:89)。

[0413] Rombutz bacillus

[0414] Amplified fragment: Regions 3958 to 4076 of the CRIB genome assembly of Rombutzella ileiformis strain, chromosome: chr1 (GenBank: LN555523.1)

[0415] Nucleic acid sequence:

[0416] GCGAAGTAGCTGATTTCACACTGTCAAGAAAAGC TTCTATCGAGTTTAAAGGCGCCCGTACCGTAAACCGACACAGGTGGGTGAGGAGAGTATCCTAAGGCCAGCGAGAGAACTATTGT (SEQ ID NO: 72)

[0417] Biomarker gene: 23S rRNA (2896 nt)

[0418] Gene product: 23S rRNA

[0419] Nucleic acid sequence:

[0420]

[0421] Gingival sulcus fungi

[0422] Amplified fragment: Regions 1611276 to 1611387 of *Eubacterium gingivalis* ATCC 35585, complete genome (GenBank: CP012068.1)

[0423] Nucleic acid sequence:

[0424] ACTGCAAACCCAAATAACAGCTTTAATAACATAAA CAATCAGAATGCTGCACCAGGCCTACAGCTACACCAGTTGCAGAGCAGGCGTCAGAGACTGACCCAGTAACACCAG (SEQ ID NO: 73)

[0425] Biomarker gene: ADJ67_07525 (624 nt)

[0426] Nucleic acid sequence:

[0427] TTATTCATTTATTGTATTCTCTGTGCCATCAACAGA (SEQ ID NO: 91)

[0428] Actinomycetes species oral taxonomy group 181

[0429] Amplified fragment: Shotgun sequence of the whole genome of strain F0379 A_sporaltaxon181-1.0_Cont73.5 of oral taxonomy group 181 of the genus Actinobacterium (GenBank: AMEW01000023.1) from region 458 to 544.

[0430] Nucleic acid sequence:

[0431] CGTGTTGGAGTGCTATTTGCGTCCGGATGCTGGTC CGCTCCTGAGTCAGTGGGTTGATCTTGAGTGGACGCCTGAGGAGGATGAATA (SEQ ID NO: 74)

[0432] Biomarker gene: HMPREF9061_01114 (545 nt)

[0433] Nucleic acid sequence:

[0434] TGGATTTTAGTGAGCGTCTTGGTCAGGTGATGCAT GATGTGTGGGGGTATGACGTGGTGGGTGATCTGGGAAAGGATGGGTATTTGGAGTTCTTTCCCACGGACACGGTCTCTGAACCGGAGGTAGTGCACTGCAAGGAGGGTTTTGTTCGCCTATTACGAGTA TGAGCGCGGCAAAATCGGTGCTCCTGTTTTTCGGTCGAGTAGTTTGCATGTGATGGAGCATTGTTTAACTCTGTGCTATGGGAATCCTCTTCGTAAAAGGTTAGGGTTTCAGCCCCTTCGTCTTGTCCG TAATTTGTCTATGCGTCCGGGATGGTCGTTGGTTCCGGTGGATTCAAAGCCTTGGCATGGGTTTTCGGGTATACGAAATAGTGAAGGTGTATTTTTCTCATGTAAGACATGGGCAGATTATCTTCTTT CGGCGTTGTCGTATGTTGTTGAGTATTCGCCTCTAGACGTGTTGGAGTGCTATTTGCGTCCGGATGCTGGTCCGCTCCTGAGTCAGTGGGTTGATCTTGAGTGGACGCCTGAGGAGGATGAATAG (SEQ ID NO: 92).

[0435] Actinomycete species S6-Spd3

[0436] Amplified fragment: Whole genome shotgun sequence of region 162210 to 162329 of contig 286 of the genus Actinobacter S6-Spd3 (GenBank: JRMV01000281.1)

[0437] Nucleic acid sequence:

[0438] TGGATTCTTTGCACTCTCCGGGGTGATTGTCATCG CGATTGTGGCGAGCGGAGCTTATGCCTTTACCGCGCATTCAACTGCCGCGGCTGTGAAGAATGTTCAAGCTCAATTGGTCGATGC (SEQ ID NO: 75)

[0439] Biomarker gene: HMPREF1627_08235 (1098 nt)

[0440] Nucleic acid sequence:

[0441] Example 3: Performance determines that the gut bacterial profile differs between CD patients and normal individuals.

[0442] Now for reference Figure 1A The graph shows the leading bacterial species identified in the study as being associated with Crohn's disease (CD). The lollipop plot in the left subplot shows the coefficient of each species with the disease calculated using MaAsLin2, adjusted for age and sex. The boxes in the middle subplot represent the phyla for each species. The bar chart in the right subplot shows the proportion of each species present in both the CD and healthy groups.

[0443] like Figure 1A As shown, MaAsLin2 analysis revealed that 46 bacterial species were negatively correlated with CD, i.e., CD-depleted species, including *Gemmelia formica*, *Eubacterium hominis*, *Broutella ovalis*, *Roseobacterium inulinum*, *Clostridium sacchariformis*, *Eubacterium rectum*, *Roseobacter faecalis*, *Bacillus sacchariformis*, *Collinsella aerogenes*, *Bacillus preibryophyte*, *Anaerobic bacillus hardruciformis*, *Trichophyton sacchariformis*, *Ruminococcus tortifolius*, *Clostridium truncatula*, *Bacteroides faecalis*, *Ruminococcus bryophyte*, *Roseobacterium enterica*, *Adlerococcus faecium*, *Alternaria putrefactiveis*, *Eubacterium* species CAG: 38, *Roseobacter humanis*, *Agastache rugosa*, *Dolphinium formate*, *Bacteroides faecium*, *Ostomyces viscera*, *Ostomyces 57_20*, *C. fecal cocci*, *Bacteroides var. var. var. var.*, *Dolphinium longiformis*, *Alternaria sarcopenia*, and *L. sacchariformis* bacteria CAG: . 83. *Oscillatoria* species CAG: 241; *Eubacterium filamentosa*, *Alternaria fenestration*, *Bacteroides foetida*, *Ackermania myxophilus*, *Concomitant cocci*, *Toxococcus butyricum*, *Alternaria inconspicuous*, *Bacteroides monomorpha*, *Bacteroides foetida*; *Eubacterium* species CAG: 274; *Desulfovibrio suis*, *Bacteroides masei*; *Clostridium* species CAG: 299. Twelve bacterial species were also found to be positively correlated with CD, i.e., CD-enriched species, namely *S. hemolyticus*, *Streptococcus lightenus*, *Actinomyces cariogenicus*, *Actinomyces* species S6-Spd3, *Actinomyces griseus*, *Lactobacillus mucosae*, *Streptococcus parahaemolyticus*, *R. mucosae*, *Actinomyces* species oral taxonomy group 181, *Veillonella spp.*, *Escherichia coli*, and *Bacteroides fragilis*.

[0444] Following further selection, 13 bacterial species (including those selected from the 46 CD-depleted species described herein and other bacterial species) were identified as potential bacterial markers for CD diagnosis and were found to be negatively correlated with CD. These CD-depleted species are: *Eubacterium tumefaciens*, *Romebuts ileum*, *Eubacterium tumefaciens*, *Eubacterium* species CAG: 274, *Dolphobicella formate*, *Anaerobic Urophyllum harzianum*, *Oscillatoria* species CAG: 241, *Oscillatoria* species 57_20, *Roseidonella enterica*, *Lactobacillus rhussi*, *Lawsonia suis*, *Ruminococcus ovalis*, and *Roseidonella inulinica*. The mean relative abundance (%) of CD-depleted species in normal individuals is shown in Table 1.2.

[0445] Five bacterial species (including those selected from the 12 CD-enriched species described in this paper and other bacterial species) were also identified as potential bacterial markers for CD diagnosis. These species were found to be positively correlated with CD, namely CD-enriched species: Actinomycete species oral taxa 181, Escherichia coli, Actinomycete species S6-Spd3, Bacteroides fragilis, and Eubacterium gingivalis. The mean relative abundance (%) of CD-enriched species in normal individuals is shown in Table 1.3. Figure 1B The diagram shows the associations between CD, sex, age, and the relative abundance of 13 CD-depleted bacterial species and 5 CD-enriched species, calculated using the machine learning model MaAsLin2. Significant correlations (FDR < 0.05) are indicated by a plus sign for positive correlations and by a minus sign for negative correlations. The false discovery rate (FDR) was calculated using Bengemini-Hochberg correction. These bacteria can be used to determine the risk or diagnose Crohn's disease in subjects.

[0446] Therefore, the bacteria listed in Tables 1.2 and 1.3 can be used in different combinations to establish assessment models to determine the risk or diagnosis of Crohn's disease in subjects, and whether microbiome recovery therapy or supplementation is needed. Relative abundance can be determined by using a set of primers for qPCR or ddPCR or by metagenomic sequencing to determine the risk or diagnosis of Crohn's disease in subjects.

[0447] The bacteria listed in Table 1.2 can be used to relieve CD symptoms in CD patients. Table 1.2: Bacterial species enriched in normal individuals compared to CD patients (CD-depleted species). Table 1.3: Bacterial species enriched in CD patients compared to normal individuals (CD-enriched species) Machine learning models based on metagenomic data, using single bacterial biomarkers or combinations of different bacterial biomarkers. Performance.

[0448] The performance of machine learning models for CD diagnosis was evaluated using single bacterial biomarkers or combinations of different bacterial biomarkers. Model performance for single bacterial biomarkers ranged from 0.391 to 0.773 (numbers 1-18 in Table 1.4). To enhance the model's disease diagnostic capabilities, different combinations of bacterial biomarkers were tested, with AUCs ranging from 0.705 to 0.936 in the discovery cohort (numbers 19-136 in Table 1.4). A combination of nine bacterial biomarkers achieved the best diagnostic performance, with an AUC of 0.936 (number 19 in Table 1.4). Table 1.4: Machine learning models based on different combinations of bacterial biomarkers using metagenomic data for CD risk Predicted performance. Performance of machine learning models based on biomarkers of nine selected bacterial species using metagenomic data.

[0449] Now for reference Figure 2A-2C The figures show the relative abundance (%) of 18 bacterial species markers identified by metagenomic sequencing (as listed in Tables 1.2 and 1.3) in CD patients and normal individuals in the Hong Kong cohort, the Hong Kong validation cohort, and the Australia (AUS) validation cohort.

[0450] To obtain models with optimal performance across different populations, different combinations of bacterial biomarkers were evaluated. Ultimately, nine selected bacterial species biomarkers (also referred to as "biomarkers") were used in the machine learning model, including *Actinomyces* species oral taxa 181, *Bacteroides fragilis*, *Escherichia coli*, *Lawsonia unsacchariformis*, *Eubacterium* species CAG: 274, *Roseidonella inulinica*, *Roseidonella enterica*, *Ruminococcus ovalis*, and *Dolphobicella formate* (Table 1.5). Figure 2A-2C As shown, the relative abundance of these nine biomarkers is consistent across the Hong Kong discovery cohort, the Hong Kong validation cohort, and the Australian validation cohort.

[0451] Now for reference Figure 3A It shows the receiver operating characteristic (ROC) curves and area under the curve (AUC) of a machine learning model (random forest model) using nine biomarkers identified through metagenomic sequencing in the training set, test set, Hong Kong validation cohort, and Australia (AUS) validation cohort. The final model using these nine biomarkers was found in the Hong Kong cohort (e.g., Figure 3AThe area under the receiver operating characteristic (ROC) curve (AUC) shown in the diagram (labeled "test set") has a value of 0.9357 (95% CI: 0.8871–0.9844; 88.33% sensitivity, 89.47% specificity, and an optimized threshold of 0.456). In the Hong Kong and Australian validation cohorts, the AUCs for these nine biomarkers were 0.8314 and 0.7264, respectively.

[0452] Now for reference Figure 3B The study shows the receiver operating characteristic (ROC) curves and area under the curve (AUC) values ​​of a machine learning model (random forest model) using nine biomarkers identified through metagenomic sequencing on three public datasets including subjects from the United States, the Netherlands, and China. The AUCs of the diagnostic model in the US, Netherlands, and China were 0.8932, 0.8636, and 0.9676, respectively, in the three public datasets.

[0453] The results showed that the machine learning model using nine biomarkers was accurate in predicting the risk of Crohn's disease in the subjects. Table 1.5: Bacterial species included in the machine learning model for CD risk prediction Determining Crohn's Disease Risk Using Different Combinations of Bacterial Biomarkers

[0454] To determine the risk of Crohn's disease or diagnose Crohn's disease in subjects, the following steps are performed: (1) Obtain a training dataset by determining the relative abundance of species selected from Table 1.2 or Table 1.3 in normal individuals and Crohn's disease patient cohorts; (2) Determine the relative abundance of these species in the subjects, whose Crohn's disease risk is to be determined; (3) Compare the relative abundance of these species in the subjects with the training data using a random forest model; and (4) Generate a decision tree from the training data using a random forest. The relative abundance is run along the decision tree and a risk score is generated. If more than 50% of the trees in the model consider the subject to have Crohn's disease, the tested subject is considered to have an increased risk of Crohn's disease. If less than 50% of the trees in the model consider the subject to be a normal individual, the tested subject is considered not to have an increased risk of Crohn's disease. Bacterial species identified by droplet digital PCR (ddPCR)

[0455] Now for reference Figures 4A-4C Using the specific primers and probes listed in Table 1.1, the levels of 6 bacterial species negatively correlated with CD and 3 bacterial species positively correlated with CD were determined by ddPCR in the subgroups of the discovery cohort (CD = 172, control = 86) (Table 1.5). The use of ddPCR data to develop machine learning models based on single bacterial biomarkers or combinations of different bacterial biomarkers. able.

[0456] The performance of machine learning models for CD diagnosis was evaluated using single bacterial biomarkers or combinations of different bacterial biomarkers. Model performance for single bacterial biomarkers ranged from 0.608 to 0.772 (numbers 1-9 in Table 1.6). To enhance the model's disease diagnostic capabilities, different combinations of bacterial biomarkers were tested, with AUCs ranging from 0.59 to 0.877 (numbers 10-95 in Table 1.6, where only combinations with AUCs greater than 0.6 were retained). A combination of nine selected bacterial biomarkers yielded superior AUCs, with a relatively balanced sensitivity and specificity, resulting in overall performance better than other combinations (number 10 in Table 1.6). Table 1.6: Machine learning models based on different combinations of bacterial biomarkers using ddPCR data for CD risk prediction Performance. Performance of machine learning models using ddPCR data

[0457] Nine selected bacterial species biomarkers were ultimately used in the machine learning model, including *Actinomyces* species oral taxa 181, *Bacteroides fragilis*, *Escherichia coli*, *Lawsonia unsacchariformis*, *Eubacterium* species CAG: 274, *Roseidobacterium inulinum*, *Roseidobacterium enterica*, *Ruminococcus ovalis*, and *Dolphobic formate* (Table 1.5). The subgroups of the cohort (CD = 172, Control = 86) were randomly assigned to a training set (CD = 131, Control = 61) and a test set (CD = 41, Control = 25).

[0458] Figure 5 The receiver operating characteristic (ROC) curves and area under the curve (AUC) of the machine learning model are shown. The AUC of the random forest model using nine biomarkers identified by droplet digital PCR is also shown. The final model using these nine biomarkers achieved an AUC of 0.999 on the training set and 0.8688 on the test set (95% CI: 0.7779–0.9596; 90.24% sensitivity, 76.00% specificity, and an optimized threshold of 0.637). Biomarkers of CD-depleted bacterial species can help deplete beneficial pathways in CD patients.

[0459] Figures 6A-6DStacked bar charts compared the relative abundance (%) of different bacterial species (i.e., Actinomycetes species oral taxa 181, Bacteroides fragilis, Broutbacterium ovale or Ruminococcus ovale, Formicotrophus dorsalis, Eubacterium species CAG: 274, Lawsonia saccharitroph, Rosette, Rosette of inulin, and other bacterial species) in the following functional pathways: starch degradation, L-tryptophan biosynthesis, thiamine pyrophosphate formation from pyridine thiamine and hydroxy thiamine (yeast), and the superpathway for L-lysine, L-threonine, and L-methionine biosynthesis II. Each stacked bar chart represents the contribution of bacterial species biomarkers and other bacteria to each functional pathway.

[0460] like Figures 6A-6D As shown, the abundance of the following functional pathways is reduced in CD patients: amino acid biosynthesis (superpathways for L-tryptophan biosynthesis, L-lysine, L-threonine, and L-methionine biosynthesis II), carbohydrate degradation (starch degradation), and cofactor, carrier, and vitamin biosynthesis (thiamine pyrophosphate formation from pyridine thiamine and hydroxy thiamine). These pathways are abundant in healthy controls. The higher abundance of these pathways in healthy controls is mainly contributed by CD-depleted species such as *Broutella ovalis*, *Roseidon ovalis*, *Dolphobicella formate*, *CAG: 274*, and *Roseidon ovalis*. Therefore, providing CD patients with effective amounts of *Broutella ovalis*, *Roseidon ovalis*, *Dolphobicella formate*, *CAG: 274*, and / or *Roseidon ovalis* can increase the abundance of these functional pathways, which is beneficial for CD treatment. Example 4: Predicting CD risk in test subjects

[0461] In this example, to determine the risk score of Crohn's disease in a subject, the following steps were performed: (1) The relative abundance of nine selected bacterial species biomarkers, as shown in Table 1.5, was measured in the subject by metagenomic sequencing or ddPCR; (2) The relative abundance data were input into a pre-trained machine learning model (random forest model) to generate a disease risk score (disease probability) for the subject; (3) If the subject's risk score (disease probability) was higher than the optimized cutoff value of 0.456 for metagenomic data or 0.637 for ddPCR data, the subject was considered to have Crohn's disease or be at risk of having Crohn's disease. Conversely, if the risk score (disease probability) was not higher than the optimized cutoff value, the individual was considered not to have Crohn's disease or be at risk of having Crohn's disease.

[0462] Table 1.7 shows the predicted risk (probability of disease) of CD in 98 subjects (including known CD patients and controls, as shown in the “Original Group”) using a machine learning model based on metagenomic data. Table 1.8 shows the predicted risk (probability of disease) of CD in 66 subjects (including known CD patients and controls, as shown in the “Original Group”) using a machine learning model based on ddPCR data.

[0463] The “Original Group” column indicates the condition of each subject as assessed by a healthcare professional through other means (e.g., colonoscopy), where “CD” refers to subjects diagnosed with Crohn’s disease, and “Control” refers to subjects without Crohn’s disease. After analyzing samples from subjects according to the disclosed methods as described in the foregoing examples, a risk score (disease probability or risk prediction) is calculated. If the risk score is higher than the optimized cutoff value, the subject is labeled “CD” in the “Test Results” column. Conversely, if the risk score is not higher than the optimized cutoff value, the subject is labeled “Control” in the “Test Results” column.

[0464] As shown in Table 1.7, in the test results using the machine learning model based on metagenomic data, 34 out of 38 subjects in the original "control" group were correctly identified as "controls," and 53 out of 60 subjects in the original "CD" group were correctly identified as "CD." As shown in Table 1.8, in the test results using the machine learning model based on ddPCR data, 19 out of 25 subjects in the original "control" group were correctly identified as "controls," and 37 out of 41 subjects in the original "CD" group were correctly identified as "CD."

[0465] The machine learning model based on metagenomic data had a sensitivity of 88.33% and a specificity of 89.47%. The machine learning model based on ddPCR data had a sensitivity of 90.24% and a specificity of 76%.

[0466] These results demonstrate that machine learning models based on metagenomic and ddPCR data are highly accurate in predicting CD risk in subjects. Table 1.7: CD risk scores of subjects when using machine learning models based on metagenomic data (disease profile) Rate). Table 1.8: CD risk score (probability of disease) of subjects when using machine learning models based on ddPCR data. Example 5: Risk prediction for CD patients with different disease activities Disease risk scores generated using a machine learning model that employs nine selected CD bacterial biomarkers can reflect metabolic disorders in CD patients.

[0467] Each subject's metabolic dysfunction score was determined by calculating the median Bray-Curtis dissimilarity to the reference control group based on metabolic functional pathways. (Now referencing...) Figure 7 The figure shows a graph illustrating the correlation between functional ecodysregulation scores and disease probabilities generated by a machine learning model based on biomarkers of nine selected CD bacterial species (according to Table 1.5). The correlation coefficients R and p-values ​​are given by Spearman correlation (…). (p < 0.05). Figure 7 As shown, the disease risk score generated by a machine learning model using nine selected CD bacterial biomarkers (listed in Table 1.5) was positively correlated with the dysfunction score. This indicates that the machine learning model can reflect the metabolic dysfunction in CD patients. A diagnostic model based on nine selected CD bacterial biomarkers can diagnose CD patients in remission.

[0468] Crohn's disease (CD) is characterized by alternating periods of disease activity, including flare-ups and remissions. The accuracy and stability of nine selected bacterial biomarkers (listed in Table 1.5) were evaluated in CD patients with varying disease activity. CD patients were categorized into those with active disease (CD: CDAI > 150) and those with inactive disease (CD: CDAI ≤ 150) using the Crohn's Disease Activity Index (CDAI).

[0469] Figures 8A-8C The results showed that, compared with healthy controls, patients with inactive Crohn's disease (CD) had a decreased number of six CD-depleted bacterial markers: *Roseidon oryzae*, *Brutella ovalis*, *Lawsonia solani*, *Roseidon oryzae*, *Dolphobicella formate*, and *CAG:274*, while compared with healthy controls, patients with inactive CD had an increased number of three CD-enriched bacterial markers: *Bacteroides fragilis*, *Escherichia coli*, and *Oral taxa 181*. P-values ​​were given by the Wilcoxon rank-sum test. p < 0.05, p < 0.01, p < 0.001, NS not significant. Figure 8DThe results showed that the disease risk scores generated by a machine learning model using nine selected CD bacterial biomarkers did not differ significantly between CD patients with active and inactive disease. Figure 8E The model achieved an AUC of 0.837 when classifying CD patients in remission against controls. This demonstrates that machine learning models can be used to determine the risk of Crohn's disease in individuals, regardless of disease activity.

[0470] Of the nine selected bacterial biomarkers, *Brutella ovalis*, *Lawsonia unsacchariformis*, and *Dolphobicia formate* showed lower levels in active CD patients compared to inactive CD patients, while *Escherichia coli* showed higher levels in active CD patients than in inactive CD patients. Therefore, these species can be used as bacterial biomarkers for monitoring disease activity in CD patients. For example, a set of biomarkers containing these species can be used to train machine learning models in cohorts of CD patients with both active and inactive disease, enabling their use as biomarkers for disease surveillance and prediction of disease onset. Example 6: Differentiating CD from other diseases A diagnostic model based on nine selected CD bacterial markers can differentiate CD from other gastrointestinal (GI) and non-GI diseases.

[0471] Given the common microbiota alterations across various diseases, validating the disease specificity of identified bacterial markers is crucial to ensure a low false-positive rate in the diagnosis of inflammatory bowel disease (IBD). To this end, several non-IBD disease datasets were evaluated, comprising subjects with gastrointestinal (GI) diseases (n = 439) and non-gastrointestinal (non-GI) diseases (n = 291). The GI diseases included colorectal cancer (CRC, n = 160), colorectal adenoma (CA, n = 162), and irritable bowel syndrome (diarrhea subtype, IBS-D, n = 117). The non-Gastrointestinal (non-GI) diseases included obesity (BMI > 28; n = 148) and cardiovascular disease (CVD; n = 143).

[0472] Now for reference Figures 9A-9B . Figure 9A The figure shows the ROC curves for classifying CD patients compared to patients with other GI diseases, including CA (n=162), IBS-D (n=117), and CRC (n=160). Figure 9AAs shown, the CD diagnostic model / machine learning model (random forest model) using nine selected bacterial biomarkers (listed in Table 1.5) identified by metagenomic sequencing can distinguish CD subjects from patients with other GI diseases, with AUROC values ​​of 0.8963, 0.8283, and 0.8078 for colorectal adenoma (CA), irritable bowel syndrome (diarrhea subtype, IBS-D), and colorectal cancer (CRC), respectively. Figure 9B This is a comparison of disease probabilities (risk scores) generated for patients with CD and those with other GI diseases using a model based on biomarkers of nine selected CD bacterial species. P-values ​​were calculated using the Wilcoxon rank-sum test (p < 0.05). p < 0.01; p < 0.001; p < 0.0001; ns, no significance). Figure 9B As shown, the disease probability (risk score) of patients with other GI diseases is significantly lower than that of patients with CD, demonstrating that the machine learning model is accurate in distinguishing between CD patients and patients with other GI diseases. Figure 9C The figure shows the ROC curves that classify CD patients from those with non-IBD GI diseases, where the CD diagnostic model distinguishes CD subjects from those with other GI diseases with a mean AUROC of 0.8459.

[0473] Now for reference Figures 10A-10B . Figure 10A The figure shows the ROC curves for classifying CD patients with other non-GI diseases, including obesity (n = 148) and CVD (n = 143), and a comparison of their disease probabilities generated by a model based on biomarkers of nine selected CD bacterial species. Figure 10A As shown, the same model can also distinguish CD subjects from patients with other non-GI diseases, with AUROCs of 0.85 and 0.918 for obesity and cardiovascular disease (CVD), respectively. Figure 10B The figure shows a comparison of disease probabilities (risk scores) generated in CD and other non-GI diseases using a model based on biomarkers of nine selected CD bacterial species. P-values ​​were calculated using the Wilcoxon rank-sum test (p < 0.05). p < 0.01; p < 0.001; p < 0.0001; ns, no significance). Figure 10B As shown, the disease probability (risk score) of patients with other non-GI diseases is significantly lower than that of CD patients, demonstrating that the machine learning model is accurate in distinguishing CD patients from patients with other non-GI diseases.

[0474] Figure 10C The figure shows the ROC curves for classifying CD patients against patients with all other non-IBD diseases (GI and non-GI diseases) (n = 730) using a machine learning model based on biomarkers of nine selected bacterial species. Figure 10C As shown, the machine learning model can distinguish CD patients from patients with all other non-IBD diseases, with an AUROC of 0.8609.

[0475] These results demonstrate that the diagnostic model encompassing nine selected CD-related biomarkers exhibits satisfactory performance in distinguishing CD from all other non-IBD diseases (i.e., GI diseases and / or other non-GI diseases).

[0476] Of the nine selected bacterial markers, *Rossiella pneumoniae* and *Brutella ovalis* were depleted in CD patients compared to other diseases, while *Bacteroides fragilis* and *Escherichia coli* were enriched in CD patients compared to other diseases. A diagnostic model based on nine selected CD bacterial markers can differentiate CD from another IBD subtype, ulcerative colitis. enteritis

[0477] Now for reference Figure 11 The figure shows the ROC curves for classifying CD patients from UC patients using a CD diagnostic model / machine learning model based on biomarkers of nine selected bacterial species (according to Table 1.5). Among the bacterial biomarkers, depletion of *Roseola inulin* and *Broutella ovalis*, and enrichment of *Bacteroides fragilis* and *Escherichia coli* are characteristic of CD. Figure 11 As shown, the CD diagnostic model can distinguish CD subjects from ulcerative colitis (UC) subjects, with an AUROC of 0.7014.

[0478] The results of differentiating CD from GI diseases and from other non-GI (non-IBD) diseases with UC based on the model using biomarkers of nine selected CD bacterial species are summarized in Table 1.9. Table 1.9: Performance of CD diagnostic models in distinguishing CD from other gastrointestinal (GI) and non-GI diseases Example 7: Methods used to identify, diagnose, prevent, or treat the following risks in an individual: [List of methods would be inserted here] Ron's disease (CD)

[0479] Now for reference Figure 12 It illustrates an example method 100 for determining, diagnosing, preventing, or treating Crohn's disease (CD) in an individual, the method involving the following steps:

[0480] Step 110: Determine the relative abundance of a set of bacterial species in a fecal sample from the individual, wherein the set of bacterial species comprises one or more bacterial species selected from the group consisting of: formic acid-producing Dorperella, picky Eubacterium, inulin-eating Rosette, enteric Rosette, ovale-forming rumenococci (ovale-forming Brontë), protruding Eubacterium, anaerobic bacilli of Hadrullus, sugar-insoluble Lawsonia, vibrio species CAG: 241, vibrio species 57_20, Lactobacillus rhamnosus, Eubacterium species CAG: 274, ileum Rombutz, Bacteroides fragilis, Escherichia coli, gingival groove Eubacterium, Actinomyces species oral taxonomy group 181 and Actinomyces species S6-Spd3;

[0481] Step 120: Compare the relative abundance of each bacterial species in the bacterial species set from the individual with the relative abundance of each bacterial species in the bacterial species set in the reference dataset to generate a risk score;

[0482] Step 130: If the risk score is higher than the cutoff value, then the individual is determined to have an increased risk of developing CD or to already have CD;

[0483] Step 140: If it is determined that the individual has an increased risk of developing CD or already has CD, then the individual may optionally be treated. Example 2

[0484] In some embodiments, a method is provided for determining, diagnosing, preventing, or treating ulcerative colitis (UC) in an individual, the method comprising the steps of: (a) determining the relative abundance of a set of bacterial species in a fecal sample from the individual, wherein the set of bacterial species comprises one or more bacterial species selected from the group consisting of: *Collagenia fecalis*, *Bacillus cryptana*, *Colinus fecalis*, *Oscillatoria* species CAG: 241, *Lawsonia cryptana*, *Butymonas toxicans*, *Clostridium* species CAG: 58, *Eubacterium* species CAG: 274, *Bacillus warfarinii*, *Clostridium tenuifolium*, *Clostridium sacchariformis*, *Gemmelia formica*, *Ostomyces viscerans*, *Ruminococcus tortifolia*, *Actinomyces* species oral taxonomy group 181, *Clostridium spiralis*, *Gimidae muscarinica*, and *Broutella henselae*; (b) The relative abundance of each bacterial species in the bacterial species set from the individual is compared with the relative abundance of each bacterial species in the bacterial species set in the reference dataset to generate a risk score; (c) if the risk score is higher than a cutoff value, the individual is determined to have an increased risk of UC or to have UC; and (d) if the individual is determined to have an increased risk of UC or to have UC, the individual is optionally treated.

[0485] In some embodiments, the bacterial species set includes one or more bacterial species selected from the group consisting of: *Colins oxytocinus*, *Clostridium truncatum*, *Ruminococcus tortifolius*, *Bacillus crypticarum*, *Gemacobacter formic acid*, *Clostridium sacchariformis*, *Alternaria putrefactiveis*, *Dolphinium longum*, *Ruminococcus brevicornu*, *Ostomyces viscerans*, *Trichophyton solani*, *Concomitant cocci*, *Adlerococcus adele*, *Roseidon*, *Brutella ovalis*, *Eubacterium rectum*, *Dolphinium formate*, *Alternaria sarcopenia*, *Bacteroides fecalis*, *Bacteroides parafecalis*, *CAG: 241* species of *Ostomyces*, *Ackermania pseudomallei*, *Alternaria fibrinolytica*, *Eubacterium hominis*, *CAG: 57-20* species of *Ostomyces humanis*, *Colins foetida*, *Bifidobacterium adolescentis*, *Roseidon*, *Toxocytotoxic butyric acid bacteria*, *Alternaria inconspicuous*, *Eubacterium* species of *CAG: 241*. 38. *Hardrophyte*, *Bacteroides foetida*, *Clostridium* species CAG: 242. *Lawsonia sucrase*, *Eubacterium* species CAG: 274. *Cercozoella fecalithii*, *Bacteroides masei*, *Roseidon*, *Clostridium* species CAG: 58. *Lactobacillus lactis*, *Bacteroides vulgaris*, *Biliobacterium warfarinii*, *Firmite* bacteria CAG: 145. *Eubacterium cleistatum*, *Enterobacter cecum*, *Twincocephala mesenteriae*, *Veillonella infantis*, *Haemophilus* species HMSC71H05, *Actinomyces* species oral taxonomy group 181. *Broutella aerogenes*, *Lactobacillus mucosa*, *Enterococcus avium*, *Veillonella atypical*, *Eubacterium gingivalis*, *Broutella hensenopyraceae*, *Rhodopseudomonas mucosae*, *Clostridium spiralis*, *Tizel*, *Veillonella septum*, and *Bacteroides fragilis*.

[0486] In some embodiments, the reference dataset is obtained by determining the relative abundance of each bacterial species in a set of bacterial species from a subject group that includes UC patients and normal individuals in a reference cohort.

[0487] In some embodiments, a machine learning model is used to perform the comparison step in step (b) to generate a risk score.

[0488] In some embodiments, the machine learning model is constructed by: (i) obtaining a reference dataset by determining the relative abundance of each bacterial species in a bacterial species set from fecal samples of a subject group including UC patients and normal individuals in a reference cohort; (ii) dividing the reference dataset into a training dataset and a test dataset; (iii) dividing the training dataset into a training subset and a validation subset; (iv) training a candidate model using the training subset and evaluating the performance of the candidate model using the validation subset; (v) optionally tuning the candidate model by adjusting the hyperparameters in the candidate model; (vi) repeating steps (iv) and optionally (v) to generate a set of candidate models and selecting the best candidate model from the set of candidate models; (vii) evaluating the performance of the best candidate model using the test dataset; (viii) optionally training a final model with the reference dataset using a set of hyperparameters corresponding to the best candidate model to generate a final model; and (ix) using the best candidate model of step (vii), or if the final model was trained using step (viii), using the final model of step (viii) as the model for step (b). The machine learning model in the comparison step generates a risk score. For clarity, the reference dataset used in step (viii) is a complete reference dataset, which includes both the training dataset and the test dataset.

[0489] In some embodiments, the machine learning model is a random forest model, and the risk score in step (b) is generated by the following steps: (b1) generating a set of decision trees using the relative abundance of the bacterial species set in the reference dataset via the random forest model; and (b2) running the relative abundance of the bacterial species set from individuals along the set of decision trees to generate a risk score.

[0490] In some embodiments, the best candidate model is the candidate model with the highest performance in the candidate model set.

[0491] In some embodiments, the performance of a candidate model is evaluated by the area under the curve (AUC) of the candidate model's receiver operating characteristic (ROC) curve. In some embodiments, the area under the curve (AUC) of the receiver operating characteristic (ROC) is also referred to as "AUROC".

[0492] In some embodiments, the cutoff value is in the range of 0.4 to 0.7.

[0493] In some embodiments, the cutoff value is 0.5.

[0494] In some embodiments, the cutoff value is determined by the following steps: (c1) calculating the Yoden index using the risk scores in the reference dataset; and (c2) determining the cutoff value based on the Yoden index.

[0495] In some embodiments, the bacterial species set includes one or more combinations of these bacterial species, wherein the combinations are selected from the group consisting of Tables 2.4 and 2.6.

[0496] In some embodiments, the bacterial species set consists essentially of the following: oral taxa 181 of the genus Actinomycetes, Clostridium spiralis, and Geminidella mesenteriae.

[0497] In some embodiments, the bacterial species set includes Clostridium sacchariformis, Clostridium truncatum, Gemmaella formica, and optionally Ruminococcus tortifolia and / or Ostomyces viscera.

[0498] In some embodiments, the bacterial species set includes *Bacillus vorinae*, *Clostridium truncatum*, *Clostridium sacchariformis*, *Gemmelia formica*, *Ostomyces viscerans*, *Ruminococcus truncatum*, *Actinomyces* species oral taxonomy group 181, *Clostridium spiralis*, *Twinococcus mesenteriae*, and *Brutella hensenopyelete*.

[0499] In some embodiments, a method is provided, wherein the bacterial species set comprises the following: *Bacillus vortex*, *Clostridium truncatum*, *Clostridium sacchariformis*, *Gemmelia formica*, *Ostomyces viscerans*, *Ruminococcus truncatum*, *Actinomyces* species oral taxonomy group 181, *Clostridium spiralis*, *Twincoccus mesenteriae*, and *Brutella hensenopyelete*.

[0500] In some embodiments, the bacterial species set includes Clostridium truncatum, Clostridium sacchariformis, Osmotherium visceratum, Gemmaietta esculenta, Clostridium spiralis, and Ruminococcus tornosa, thereby the method identifies an individual's risk of developing UC relative to other non-IBD diseases.

[0501] In some embodiments, other non-IBD conditions include colorectal cancer, colorectal adenoma, irritable bowel syndrome (diarrhea subtype), obesity, and cardiovascular disease.

[0502] In some embodiments, the relative abundance of the bacterial species set is determined by metagenomic sequencing or droplet digital PCR (ddPCR).

[0503] In some embodiments, the relative abundance of the bacterial species set is determined by metagenomic sequencing, and the bacterial species set comprises a combination of one or more bacterial species selected from Table 2.4.

[0504] In some embodiments, the relative abundance of the bacterial species set is determined by ddPCR, and the bacterial species set comprises a combination of one or more bacterial species selected from Table 2.6.

[0505] In some embodiments, ddPCR is performed using one or more primer sets selected from Table 2.1.

[0506] In some embodiments, if an individual is determined to have an increased risk of developing UC or already has UC, the individual is treated with an appropriate UC treatment selected from the group consisting of: therapeutic therapy, fecal microbiota transplantation (FMT), intra-intestinal microbiota transplantation (IMT), lifestyle and dietary changes, and surgical intervention. In some embodiments, the therapeutic therapy is selected from the group consisting of: 5-aminosalicylates (5-ASA, such as mesalazine, sulfasalazine); steroids (such as budesonide, prednisolone); thiopurines (such as azathioprine, mercaptopurine, methotrexate); TNF inhibitors (such as infliximab, adalimumab, and sertozumab); anti-integrins (such as vedolizumab); and anti-interleukin (IL) 12 / 23 (such as ustekinumab).

[0507] In some embodiments, 5-aminosalicylate is mesalazine or sulfasalazine; steroid is budesonide or prednisolone; thiopurine is azathioprine, mercaptopurine, or methotrexate; TNF inhibitor is infliximab, adalimumab, or cetuzumab; anti-integrin is vedolizumab; and anti-interleukin (IL) 12 / 23 is ustekinumab.

[0508] In some embodiments, UC is in a mitigation phase.

[0509] In some embodiments, a method is provided for identifying a fecal sample with an altered ulcerative colitis microbiome, the method comprising the steps of: (a) determining the relative abundance of a set of bacterial species in the fecal sample, wherein the set of bacterial species comprises one or more bacterial species selected from the group consisting of: *Collagenia fecalis*, *Bacillus spp.*, *Colinus spp.*, *Oscillatoria* species CAG: 241, *Lawsonia spp.*, *Toxococcus butyricum*, *Clostridium* species CAG: 58, *Eubacterium* species CAG: 274, *Bacillus warfarinii*, *Clostridium spp.*, *Clostridium spp.*, *Clostridium spp.*, *Gemmelia formica*, *Ostomyces viscerans*, *Ruminococcus tortifolius*, *Actinomyces* species oral taxonomy group 181, *Clostridium spiralis*, *Gimidae muscarinica*, and *Broutella henselae*; (b) comparing the relative abundance of each bacterial species in the set of bacterial species from the individual with the relative abundance of each bacterial species in the set of bacterial species in a reference dataset to generate a risk score; and (c) If the risk score is higher than the cutoff value, the fecal sample is determined to have an altered UC microbiome.

[0510] In some embodiments, a machine learning model is used to perform the comparison step in step (b) to generate a risk score.

[0511] In some embodiments, the machine learning model is constructed by: (i) obtaining a reference dataset by determining the relative abundance of each bacterial species in a bacterial species set from fecal samples from a subject group including UC patients and normal individuals in a reference cohort; (ii) dividing the reference dataset into a training dataset and a test dataset; (iii) dividing the training dataset into a training subset and a validation subset; (iv) training a candidate model using the training subset and evaluating the performance of the candidate model using the validation subset; (v) optionally tuning the candidate model by adjusting the hyperparameters in the candidate model; (vi) repeating steps (iv) and optionally (v) to generate a set of candidate models and selecting the best candidate model from the set of candidate models; (vii) evaluating the performance of the best candidate model using the test dataset; (viii) optionally training a final model with the reference dataset using a set of hyperparameters corresponding to the best candidate model to generate a final model; and (ix) using the best candidate model from step (vii), or if the final model was trained using step (viii), using the final model from step (viii) as the model for step (b). A machine learning model is used for comparison steps to generate risk scores.

[0512] In some embodiments, the machine learning model is a random forest model, and the risk score in step (b) is generated by the following steps: (b1) generating a set of decision trees using the relative abundance of the bacterial species set in the reference dataset via the random forest model; and (b2) running the relative abundance of the bacterial species set from individuals along the set of decision trees to generate a risk score. In some embodiments, the risk score is generated by averaging predictions across decision trees or by majority voting on the predicted class labels across decision trees, wherein the proportion of votes in each class across the set is the predicted probability (i.e., the risk score).

[0513] In some embodiments, the best candidate model is the candidate model with the highest performance in the candidate model set.

[0514] In some embodiments, the performance of a candidate model is evaluated by the area under the curve (AUC) of the receiver operating characteristic (ROC) curve of the candidate model.

[0515] In some embodiments, the cutoff value is in the range of 0.4 to 0.7.

[0516] In some embodiments, the cutoff value is 0.5.

[0517] In some embodiments, the cutoff value is determined by the following steps: (c1) calculating the Yoden index using the risk scores in the reference dataset; and (c2) determining the cutoff value based on the Yoden index.

[0518] In some embodiments, the bacterial species set includes one or more combinations of these bacterial species, wherein the combinations are selected from the group consisting of Tables 2.4 and 2.6.

[0519] In some embodiments, the bacterial species set consists essentially of the following: oral taxa 181 of the genus Actinomycetes, Clostridium spiralis, and Geminidella mesenteriae.

[0520] In some embodiments, the bacterial species set includes Clostridium sacchariformis, Clostridium truncatum, Gemmaella formica, and optionally Ruminococcus tortifolia and / or Ostomyces viscera.

[0521] In some embodiments, the bacterial species set includes *Bacillus vorinae*, *Clostridium truncatum*, *Clostridium sacchariformis*, *Gemmelia formica*, *Ostomyces viscerans*, *Ruminococcus truncatum*, *Actinomyces* species oral taxonomy group 181, *Clostridium spiralis*, *Twinococcus mesenteriae*, and *Brutella hensenopyelete*.

[0522] In some embodiments, the bacterial species set comprises the following: *Bacillus vorinae*, *Clostridium truncatum*, *Clostridium sacchariformis*, *Gemmelia formica*, *Ostomyces viscerans*, *Ruminococcus truncatum*, *Actinomyces* species oral taxonomy group 181, *Clostridium spiralis*, *Twincoccus mesenteriae*, and *Brutella hensenopyelete*.

[0523] In some embodiments, the relative abundance of the bacterial species set is determined by metagenomic sequencing or droplet digital PCR (ddPCR).

[0524] In some embodiments, the relative abundance of the bacterial species set is determined by metagenomic sequencing, and the bacterial species set comprises a combination of one or more bacterial species selected from Table 2.4.

[0525] In some embodiments, the relative abundance of the bacterial species set is determined by ddPCR, and the bacterial species set comprises a combination of one or more bacterial species selected from Table 2.6.

[0526] In some embodiments, ddPCR is performed using one or more primer sets selected from Table 2.1.

[0527] In some embodiments, a method is provided to increase amino acid biosynthesis in an individual, wherein the amino acid is L-arginine, L-ornithine, or L-valine, the method comprising providing the individual with an effective amount of one or more bacterial species selected from the group consisting of: Clostridium truncatum, Clostridium sacchariformis, Gemmaieux formica, and Ruminococcus tortifolia.

[0528] In some embodiments, a method is provided to increase amino acid biosynthesis in cells, wherein the amino acid is L-arginine, L-ornithine, or L-valine, the method comprising providing the cells with an effective amount of one or more bacterial species selected from the group consisting of: Clostridium perfringens, Clostridium sacchariformis, Gemmaieux formica, and Ruminococcus tortifolia.

[0529] In some embodiments, a method is provided to increase nucleoside and nucleotide biosynthesis in an individual, the method comprising providing the individual with an effective amount of one or more bacterial species selected from the group consisting of: Clostridium truncatum, Clostridium sacchariformis, Gemma formica, and Ruminococcus tortifolia.

[0530] In some embodiments, a method is provided to increase nucleoside and nucleotide biosynthesis in cells, the method comprising providing the cells with an effective amount of one or more bacterial species selected from the group consisting of: Clostridium truncatum, Clostridium sacchariformis, Gemmaella formica, and Ruminococcus tortifolia.

[0531] In some embodiments, the cell is part of a living organism.

[0532] In some embodiments, a composition is provided comprising Clostridium sacchariformis, Clostridium truncatum, and Gemmaella formica.

[0533] In some embodiments, the composition further comprises *Ruminococcus twisteri*.

[0534] In some embodiments, the composition further comprises one or more bacterial species selected from the group consisting of: *Collagenia fecalis*, *Bacillus crypticarum*, *Collinus faecium*, *Oscillatoria* species CAG: 241, *Lawsonia crypticarum*, *Toxococcus butyricum*, *Clostridium* species CAG: 58, *Eubacterium* species CAG: 274, and *Bacillus warfarinus*.

[0535] In some embodiments, a kit is provided for determining the risk of or diagnosing ulcerative colitis (UC) in an individual, the kit comprising reagents for detecting a set of bacterial species selected from the group consisting of: *Collagenia fecalis*, *Bacillus cryptica*, *Colinus feces*, *Oscillatoria* species CAG:241, *Lawsonia cryptica*, *Toxococcus butyricum*, *Clostridium* species CAG:58, *Eubacterium* species CAG:274, *Bacillus warwickii*, *Clostridium tenuifolium*, *Clostridium sacchariformis*, *Gemmelia formica*, *Ostomyces viscerans*, *Ruminococcus tortifolia*, *Actinomyces* species oral taxonomy group 181, *Clostridium spiralis*, *Twinia measles*, and *Brutella henselae*.

[0536] In some embodiments, the reagent includes one or more primer sets, wherein each primer set is configured to amplify nucleic acid sequences of bacterial species belonging to a set of bacterial species.

[0537] In some embodiments, one or more primer sets are selected from Table 2.1.

[0538] In some embodiments, amplification is PCR.

[0539] In some embodiments, the PCR is ddPCR, and the reagent further comprises one or more probes selected from Table 2.1.

[0540] In some embodiments, the bacterial species set includes one or more combinations of these bacterial species, wherein the combinations are selected from the group consisting of: Table 2.6.

[0541] In some embodiments, the bacterial species set consists essentially of the following: oral taxa 181 of the genus Actinomycetes, Clostridium spiralis, and Geminidella mesenteriae.

[0542] In some embodiments, the bacterial species set includes Clostridium sacchariformis, Clostridium truncatum, Gemmaella formica, and optionally Ruminococcus tortifolia and / or Ostomyces viscera.

[0543] In some embodiments, the bacterial species set includes *Bacillus vorinae*, *Clostridium truncatum*, *Clostridium sacchariformis*, *Gemmelia formica*, *Ostomyces viscerans*, *Ruminococcus truncatum*, *Actinomyces* species oral taxonomy group 181, *Clostridium spiralis*, *Twinococcus mesenteriae*, and *Brutella hensenopyelete*.

[0544] In some embodiments, the bacterial species set comprises the following: *Bacillus vorinae*, *Clostridium truncatum*, *Clostridium sacchariformis*, *Gemmelia formica*, *Ostomyces viscerans*, *Ruminococcus truncatum*, *Actinomyces* species oral taxonomy group 181, *Clostridium spiralis*, *Twincoccus mesenteriae*, and *Brutella hensenopyelete*.

[0545] In some embodiments, a computer program product is provided for determining the risk of or diagnosing ulcerative colitis (UC) in an individual, wherein the computer program product includes a computer-readable medium encoded with computer-executable code, wherein the computer-executable code is configured to perform the following steps: a) determining the relative abundance of a set of bacterial species in a fecal sample from the individual, wherein the set of bacterial species comprises one or more bacterial species selected from the group consisting of: *Coalabacterium feces*, *Bacillus spp.*, *Clinobacterium feces*, *Oscillatoria* species CAG: 241, *Lawsonia spp.*, *Toxococcus butyricum*, *Clostridium* species CAG: 58, *Eubacterium* species CAG: 274, *Bacillus warfarinii*, *Clostridium spp.*, *Clostridium spp.*, *Clostridium spp.*, *Gemmelia formica*, *Ostomyces viscerans*, *Ruminococcus tortifolius*, *Actinomyces* species oral taxonomy group 181, *Clostridium spiralis*, *Twinococcus mesenteriae*, and *Broutella henselae*; b) The relative abundance of each bacterial species in the bacterial species set from the individual is compared with the relative abundance of each bacterial species in the bacterial species set in the reference dataset to generate a risk score; and c) if the risk score is higher than a cutoff value, the individual is determined to have an increased risk of UC or to have UC.

[0546] In some embodiments, the reference dataset is obtained by determining the relative abundance of each bacterial species in the bacterial species set from fecal samples from a subject group comprising UC patients and normal individuals in a reference cohort.

[0547] In some embodiments, a machine learning model is used to perform the comparison step in step (b) to generate a risk score.

[0548] In some embodiments, the machine learning model is constructed by: i) obtaining a reference dataset by determining the relative abundance of each bacterial species in a bacterial species set from fecal samples of a subject group comprising UC patients and normal individuals in a reference cohort; ii) dividing the reference dataset into a training dataset and a test dataset; iii) dividing the training dataset into a training subset and a validation subset; iv) training a candidate model using the training subset and evaluating the performance of the candidate model using the validation subset; v) optionally tuning the candidate model by adjusting the hyperparameters in the candidate model; vi) repeating steps (iv) and optionally (v) to generate a set of candidate models and selecting the best candidate model from the set of candidate models; vii) evaluating the performance of the best candidate model using the test dataset; viii) optionally training a final model with the reference dataset using a set of hyperparameters corresponding to the best candidate model to generate a final model; and ix) using the best candidate model from step (vii), or if the final model was trained using step (viii), using the final model from step (viii) as the model for step (b). The machine learning model in the comparison step is used to generate the risk score.

[0549] In some embodiments, the machine learning model is a random forest model, and the risk score in step (b) is generated by the following steps: (b1) generating a set of decision trees using the relative abundance of the bacterial species set in the reference dataset via the random forest model; and (b2) running the relative abundance of the bacterial species set from individuals along the set of decision trees to generate a risk score.

[0550] In some embodiments, the best candidate model is the candidate model with the highest performance in the candidate model set.

[0551] In some embodiments, the performance of a candidate model is evaluated by the area under the curve (AUC) of the receiver operating characteristic (ROC) curve of the candidate model.

[0552] In some embodiments, the cutoff value is in the range of 0.4-0.7.

[0553] In some embodiments, the cutoff value is 0.5.

[0554] In some embodiments, the cutoff value is determined by the following steps: (c1) calculating the Yoden index using the risk scores in the reference dataset; and (c2) determining the cutoff value based on the Yoden index.

[0555] In some embodiments, a method is provided for determining, diagnosing, preventing, or treating ulcerative colitis (UC) in an individual, the method comprising the steps of: a) determining the relative abundance of a set of bacterial species in a fecal sample from the individual, wherein the set of bacterial species comprises one or more bacterial species selected from the group consisting of: *Colinus oxytetracyclis*, *Clostridium truncatella*, *Ruminococcus tortifolius*, *Bacillus sacchariformis*, *Gemacobacterium formica*, *Clostridium sacchariformis*, *Alternaria putrefactiveis*, *Dolphinium longiformis*, *Ruminococcus brevicornu*, *Ostomyces viscerans*, *Trichophyton floccosum*, *Concomitant cocci*, *Adlerococcus adrenocorticotropicus*, *Roseidonella inulinophila*, *Brutella ovalis*, *Eubacterium rectum*, *Dolphinium formica*, *Alternaria sarcopenia*, *Bacteroides fecalithana*, *Bacteroides fecalithana*, and *CAG* species. 241. *Ackermania pseudomallei*, *Alternaria fibrinolytica*, *Eubacterium hopterii*, *Oscillatoria* species 57_20, *Rossiella humanis*, *Koalacus fecal*, *Collinus faecium*, *Bifidobacterium adolescentis*, *Rossiella enterica*, *Toxic Butyrosporum*, *Alternaria unobservatum*, *Eubacterium* species CAG: 38, *Anaerobic Bacillus hardrovirus*, *Bacteroides fecal*, *Clostridium* species CAG: 242, *Lawsonia suisinensis*, *Eubacterium* species CAG: 274, *Cercozoella fecalithii*, *Bacteroides masei*, *Rossiella fecal*, *Clostridium* species CAG: 58, *Russobacterium lactis*, *Bacteroides vulgaris*, *Biliobacterium warfarinii*, Firmicutes bacteria CAG: 145. *Bacillus cereus*, *Enterobacter cecum*, *Twincocephalus measlesus*, *Veillonella infantis*, *Haemophilus* species HMSC71H05, *Actinomyces* species oral taxa 181, *Brutella aerogenes*, *Lactobacillus mucosae*, *Enterococcus avium*, *Veillonella atypical*, *Bacillus gingivalis*, *Brutella henseniella*, *Rhodopseudomonas myxobolus*, *Clostridium spiralis*, *Tizel*, *Veillonella septum*, and *Bacteroides fragilis*; b) compare the relative abundance of each bacterial species in the bacterial species set from the individual with the relative abundance of each bacterial species in the bacterial species set in the reference dataset to generate a risk score; c) if the risk score is higher than a cutoff value, determine that the individual has an increased risk of developing UC or currently has UC; and d) if the individual is determined to have an increased risk of developing UC or currently has UC, optionally treat the individual.

[0556] In some embodiments, a method is provided for monitoring ulcerative colitis (UC) disease activity in an individual, the method comprising the steps of: a) determining the relative abundance of a set of bacterial species in a fecal sample from the individual, wherein the set of bacterial species comprises one or more bacterial species selected from the group consisting of: Clostridium truncatum, Clostridium sacchariformis, Osmotherium visceratum, Gemmaella formica, Actinomyces species oral taxonomy 181, and Clostridium spiralis; b) comparing the relative abundance of each bacterial species in the set of bacterial species from the individual with the relative abundance of each bacterial species in the set of bacterial species in a reference dataset to generate a risk score; and c) determining that the individual has increased disease activity if the risk score is higher than a cutoff value.

[0557] In some embodiments, the reference dataset is obtained by determining the relative abundance of each bacterial species in a set of bacterial species from a subject group that includes patients with active UC and inactive UC patients in a reference cohort.

[0558] Risk assessment or diagnosis of ulcerative colitis

[0559] 1. Workflow

[0560] In one embodiment, to determine the risk of ulcerative colitis in an individual or whether the individual has ulcerative colitis, the following steps are performed: 1. Obtain stool samples from the individual and determine the level / relative abundance of one or more bacterial species selected from Tables 2.2 and 2.3 (e.g., combinations specified in Tables 2.4 or 5). 2. Obtain stool samples from a reference cohort comprising subjects with or without ulcerative colitis (UC) and determine the level / relative abundance of the same bacterial species as in step (1). 3. Compare the level / relative abundance of the bacterial species obtained in step (1) with the level / relative abundance of the bacterial species obtained in step (2).

[0561] As used in this paper, a reference cohort refers to a group of individuals with ulcerative colitis and those without (healthy controls or normal individuals) who provide samples (e.g., stool samples) to determine the levels of preselected bacterial species in such samples, where the diagnosis of ulcerative colitis is made by endoscopy, radiology, and histology; and the controls are individuals without ulcerative colitis or without any signs or symptoms of gastrointestinal disease. For a proper reference cohort to be established, each of the ulcerative colitis group and the control group must include a sufficient number of individuals with and without ulcerative colitis (e.g., at least 10, 12, 15, 20, 24, or more individuals) to provide samples for determining one or more average levels of one or more preselected bacterial species.

[0562] 2. Determine the levels of bacterial markers.

[0563] In one embodiment, a method is provided that measures the level or amount of characteristic DNA or RNA of one or more bacterial species found in a human fecal sample as a means of diagnosing or assessing the risk of ulcerative colitis. Therefore, the first step of this method is to obtain a fecal sample from a test subject and extract microbial DNA or RNA from the sample.

[0564] Fecal sample acquisition and preparation

[0565] The exemplary methods disclosed herein are used to obtain stool samples from individuals undergoing testing or monitoring for ulcerative colitis. Stool samples from individuals can be easily collected, whether in a clinic or at the patient's home. An appropriate amount of stool is collected and stored according to standard procedures for further preparation. According to the exemplary methods, established techniques can be used to analyze bacterial DNA or RNA found in the patient's stool samples. Methods for preparing stool samples for nucleic acid extraction are well known to those skilled in the art.

[0566] DNA and RNA extraction and quantification

[0567] The methods used to extract DNA from biological samples are well-known and standard practice in the field of molecular biology. RNA contamination should be eliminated to avoid interfering with DNA analysis.

[0568] Similarly, there are many methods for extracting DNA from biological samples. The most common DNA extraction methods are mechanical, chemical, and enzymatic lysis, precipitation, purification, and concentration. Other specific methods for DNA extraction include phenol-chloroform extraction, alcohol precipitation, or silica-based purification. Commercial kits, such as the MO BIO PowerSoil DNA Isolation Kit (MO Biolabs, Carlsbad, CA, USA), the QIAamp DNA Stool Mini Kit (Qiagen, Hilden, Germany), and the Maxwell® RSC PureFood GMO and Identification Kit (Promega), can also be used to obtain DNA from biological samples of test subjects. The quality and concentration of the extracted DNA will be further evaluated.

[0569] PCR-based quantitative determination of bacteria

[0570] Once DNA is extracted from a sample, the amount of predetermined bacterial DNA (e.g., 16S rDNA or DNA encoded by bacterial genes specific to the bacterial species) is quantified. In some embodiments, the method for determining DNA levels is based on amplification methods, such as polymerase chain reaction (PCR) such as real-time quantitative PCR or droplet digital PCR (ddPCR), for DNA quantification analysis.

[0571] General methods for PCR are well known in the art and will not be described in detail here. PCR reagents and protocols are also available from commercial suppliers such as Roche Molecular Systems.

[0572] PCR is most commonly performed as an automated process using thermostable enzymes. In this process, the temperature of the reaction mixture is automatically cyclicated through the denaturation zone, primer annealing zone, and extension reaction zone. Machines specifically modified for this purpose are commercially available.

[0573] Real-time quantitative PCR (a semi-quantitative PCR method) measures fluorescence after each cycle, and the intensity of the fluorescence signal reflects the instantaneous amount of DNA amplicon in the sample at that specific time.

[0574] Droplet-based digital PCR (ddPCR) is an improvement on the conventional polymerase chain reaction (PCR) method. In ddPCR, DNA / RNA is randomly encapsulated within droplets that serve as the reaction chamber. It provides more sensitive nucleic acid detection and absolute quantification.

[0575] Although PCR amplification of target bacterial DNA or RNA is typically used in implementing some of the embodiments disclosed herein, those skilled in the art will recognize that amplification of these DNA or RNA species in a sample can be accomplished by any known method, such as ligase chain reaction (LCR), transcription-mediated amplification and self-sustaining sequence replication or nucleic acid sequence-based amplification (NASBA), each of which provides sufficient amplification. Recently developed branched DNA techniques can also be used to quantitatively determine the amount of DNA or mRNA in a sample.

[0576] Bacterial Quantitative Determination Based on Metagenomics

[0577] Metagenomic datasets can include sequencing data generated from the same standardized protocol, encompassing steps from fecal DNA extraction to sequencing, raw data quality filtering, host read decontamination, and microbiome interpretation. Various techniques, such as shotgun metagenomic sequencing, can be used to characterize the species in a sample. In shotgun metagenomic sequencing, DNA is obtained from heterogeneous cellular samples and fragmented into DNA fragments that can be compared with the genomes of multiple microbial species to identify the species in the sample.

[0578] 3. Generation of Disease Risk Score (Disease Probability)

[0579] Building machine learning models

[0580] In one embodiment, a machine learning model, such as a random forest model, can be used to generate disease risk scores (disease probabilities). Random forest is a supervised learning approach for machine learning, used for classification and regression. It is a supervised machine learning algorithm that improves the prediction accuracy of a dataset by averaging or making a final decision based on the majority votes of many decision trees applied to different subsets of the dataset.

[0581] In this regard, in one embodiment, the method includes constructing a machine learning model by: (1) obtaining a reference dataset in a reference queue by determining the relative abundance of bacterial species selected from Tables 2.2 and 2.3 (e.g., combinations specified in Tables 2.4 or 2.5) in a fecal sample; (2) splitting the data into a training set and a test set; (3) splitting the data into a training subset and a validation subset under the training dataset, training a candidate model using the training subset, and evaluating the performance of the candidate model using the validation subset; (4) tuning the candidate model by using different combinations of hyperparameters; (5) repeating steps (3)-(4) and selecting the best candidate model with the best performance; and (6) evaluating the candidate model using the test set; and (7) as an optional step, training a final model using the entire reference dataset with the hyperparameter combination of the best candidate model to obtain the final model; (8) using the best candidate model from step (6), or if step (viii) is used to train the final model, using the final model from step (7) as the machine learning model.

[0582] After constructing the machine learning model, the model can be deployed to generate a risk score (disease probability) for a subject whose risk of ulcerative colitis is to be determined. In one embodiment, the method includes: (1) determining the relative abundance of the corresponding bacterial species in the individual for whom a diagnosis of ulcerative colitis is to be determined (the same as that used to construct the machine learning model). In one embodiment, the method includes: (2) inputting the relative abundance of these species from the individual obtained in step (1) into the constructed machine learning model to generate a risk score (disease probability); if a random forest model is used, the relative abundance of the species listed in Tables 2.2 and 2.3 obtained from the individual in step (1) is run along a decision tree in the random forest model; (3) determining that the individual has a risk of ulcerative colitis or has ulcerative colitis when the risk score is greater than a cutoff value, and determining that the individual has no risk of ulcerative colitis or does not have ulcerative colitis when the risk score is less than a cutoff value.

[0583] In some embodiments, the cutoff value can be determined as follows.

[0584] Determination of cutoff value

[0585] After obtaining a disease risk score (disease probability) for each subject, a default cutoff value of 0.5 or an optimized cutoff value can be used to determine whether the subject is currently suffering from ulcerative colitis or is at risk of developing ulcerative colitis.

[0586] In some embodiments, an optimized cutoff value can be calculated based on the Youden index using risk score data from a reference cohort. In some embodiments, the risk score data is obtained by generating a risk score for each bacterial species by inputting the relative abundance of bacterial species in the test set of the reference cohort into a machine learning model. The Youden index can integrate sensitivity and specificity information, and by using Youden index analysis, an optimal cutoff value that provides the best trade-off between sensitivity and specificity can be obtained. The Youden index (Y) can be calculated using the following formula:

[0587]

[0588] If a subject's risk score (disease probability) is higher than the optimized cutoff value, the subject is considered to have ulcerative colitis or be at risk of developing ulcerative colitis. Conversely, if the risk score (disease probability) is not higher than the optimized cutoff value, the individual is considered not to have ulcerative colitis or be at risk of developing ulcerative colitis.

[0589] Example methods for determining UC risk in subjects

[0590] In one embodiment, to determine a risk score for ulcerative colitis (UC) in a subject, the following steps are performed: (1) measuring the relative abundance of a selected set of bacterial species in the subject; (2) inputting the relative abundance data into a pre-trained random forest model to generate a disease risk score (disease probability) for the subject; and (3) if the subject's risk score (disease probability) is higher than an optimized cutoff value, the subject is considered to have ulcerative colitis or be at risk of having ulcerative colitis. Conversely, if the risk score (disease probability) is not higher than an optimized cutoff value, the individual is considered not to have ulcerative colitis or be at risk of having ulcerative colitis. Example 1: Methods (sample collection, DNA extraction, and sequencing)

[0591] Cohort description and study subjects

[0592] In this study, shotgun metagenomic profiling of fecal microbiomes from two different cohorts was analyzed to discover and validate diagnostic models constructed using biomarkers of ulcerative colitis-specific (UC-specific) bacterial species. In the discovery cohort, 323 Chinese participants (aged 21–87 years) were recruited, including 205 patients with ulcerative colitis (UC) and 118 healthy individuals. Males comprised 44.3%. In the validation cohort, 278 Chinese participants (aged 20–88 years) were recruited, including 139 patients with UC and 139 healthy individuals. Males comprised 54.3%. Patients were included if they were 18 years of age or older and had a diagnosis of UC as defined by endoscopy, radiology, and histology; and were on stable medication. Patients were excluded if they had: infection with intestinal pathogens; short bowel syndrome; or severe liver, kidney, endocrine, respiratory, neurological, or cardiovascular disease. Healthy individuals without serious diseases (such as inflammatory bowel disease, cancer, or advanced adenoma) were included as controls. All participants consented to donate stool samples and participated in a questionnaire, for which written informed consent was obtained. Stool samples from study participants were stored at -80°C for downstream microbiome analysis.

[0593] In addition, a total of 248 subjects from three public datasets were downloaded for model validation, including 87 subjects from the United States (53 UC patients and 34 normal individuals, aged between 20 and 82 years), 45 subjects from the Netherlands (23 UC patients and 22 normal controls, aged between 19 and 80 years), and 40 subjects from China (25 UC patients and 15 normal controls).

[0594] Fecal DNA extraction and DNA sequencing

[0595] Fecal bacterial DNA was extracted using a modified standard protocol (Maxwell® RSC PureFood GMO and Identification Kit, Promega) to improve DNA yield. Approximately 100 mg of each fecal sample was pretreated: the sample was suspended in 1 ml ddH2O and precipitated by centrifugation at 13,000 × g for 1 min. The washed sample was thoroughly mixed with 800 μL TE buffer (pH 7.5), 16 μL β-mercaptoethanol, and 250 U lysozyme, and digested at 37°C for 90 min, followed by centrifugation at 13,000 × g for 3 min to precipitate the remaining DNA.

[0596] After pretreatment, the precipitate was resuspended in 800 μL of CTAB buffer (Maxwell® RSC PureFoodGMO and Identification Kit, according to the manufacturer's instructions) and thoroughly mixed. The samples were heated at 95°C for 5 minutes and then cooled, followed by vortexing at 2850 rpm for 15 minutes using 0.5 mm and 0.1 mm beads to release nucleic acids from the samples. Subsequently, 40 μL of proteinase K and 20 μL of ribonuclease A were added, and the nucleic acids were digested at 70°C for 10 minutes. Finally, the supernatant was obtained after centrifugation at 13,000 × g for 5 minutes and transferred to a Maxwell® RSC instrument for DNA extraction. The extracted fecal DNA was used for in-house (MagIC, Hong Kong) or outsourced (Novogene, Beijing, China) ultra-deep metagenomic sequencing using an Ilumina Novaseq 6000.

[0597] Quality control of raw sequences

[0598] The original sequence reads were pruned using Trimmomatic (v0.39). Non-human reads were then separated from contaminated host reads. The steps to obtain clean reads included: 1) removing adaptors; 2) scanning the reads using a 4-base-wide sliding window and removing reads when the average quality per base dropped below 20; and 3) reducing the read length to below 50 bases. The pruned sequence reads were mapped to the human genome (reference database: hg37decv0.1) using KneadData (v0.10.0) to remove host-derived reads. The two reads at their paired ends were tandemly joined. Example 2: Methods for determining the composition of the human gut microbiota and the differentially expressed bacterial species between UC patients and healthy individuals.

[0599] Bacterial microbiome analysis

[0600] Bacterial community composition profiles were analyzed using metagenomically pruned reads via MetaPhlAn3 (v3.0.13). Reads were mapped to clade-specific marker genes and species pangenomes were annotated using Bowtie2 (v2.2.4.2). The output table contains bacterial species and their relative abundance at different levels (from kingdom to strain). The data were analyzed in R v3.6.1 using ggpubr (v0.2) and phyloseq (v1.24.2). Human gut bacterial composition and selected bacterial species were compared between UC patients and healthy individuals using multivariate linear model association analysis of the microbiome (MaAsLin2).

[0601] Machine learning models

[0602] MaAsLin2 is used for the identification of discriminative features. Random Forest (RF) is chosen to build a predictive model of UC patients compared to normal controls using selected fecal microbiota, as RF has superior performance in binary feature classification. Random Forest is one of the approaches to building predictive models in metagenomic data analysis. As a widely used ensemble learning algorithm, Random Forest consists of a series of Classification and Regression Trees (CART) to form a powerful classifier. A subset of data randomly sampled from the original dataset with replacement is called bootstrap sampling, which is suitable for building trees. When the training dataset for the current tree is plotted using the bootstrap method, Observations are excluded from the overall dataset. In the case of infinite N, 36.8% of the data does not appear in the training samples, known as out-of-bag (OOB) observations, which will not be used to build the tree. Furthermore, additional randomness is introduced into the random forest because each decision tree splits nodes based on a random subset of features selected from the overall features. In each iteration, nodes are split to generate a tree using features with the smallest Gini coefficient (the Gini coefficient is used to evaluate the purity of a node). Using different subsets of data and features, the algorithm is able to train different trees and obtain the final classification by averaging or majority voting the results from the tree models.

[0603] A total of 205 UC patients and 118 healthy controls were included as the discovery cohort for modeling. Selected species identified by MaAsLin2 were used as input for random forest model construction. Model performance was evaluated using 5-fold cross-validation. For binary classifiers, the area under the receiver operating characteristic (ROC) curve (AUC) was used to evaluate model performance. The parameters of the model construction were tuned to achieve optimal accuracy and κ. These analyses were performed using the R packages MaAsLin2 v2_1.4.0, randomForest v4.6-14, and pROC v1.18.2.

[0604] Droplet digital PCR (ddPCR)

[0605] Primer and probe design for ten selected bacterial species

[0606] Specific gene sequences for each bacterium were downloaded from the MetaPhlAn3 database, and the specificity of each bacterial gene sequence was confirmed by performing a BLAST procedure using a public database such as GenBank. Based on the gene sequences, primers and probes were designed in Primer3Plus (https: / / www.primer3plus.com / index.html) by evaluating the Tm value, GC content, and possible secondary structure of the primers and probes. The primer and probe sequences used for the 16S rDNA internal control were identical to those reported in previous studies. Primers and probes were synthesized by BGI Genomics (Hong Kong) Co., Ltd.

[0607] Detection of selected bacterial species using droplet digital PCR (ddPCR).

[0608] Microdroplet digital PCR reactions were designed to detect selected bacterial species (listed in Tables 2.2 and 2.3). The first reaction was designed to detect *Brutella hensenii* (180 nM probe), *Clostridium tenuifolium* (400 nM probe), *Actinomyces* species oral taxa 181 (180 nM probe), and *Gastrococcus mesenteriae* (450 nM probe). The second reaction was designed to detect *Clostridium spiralis* (120 nM probe), *Clostridium sacchariformis* (480 nM probe), *Ostomyces viscerans* (200 nM probe), and *Gastromyces formica* (450 nM probe). The third reaction was designed to detect *Biliobacterium warfarinii* (120 nM probe), *Ruminococcus tortifolia* (480 nM probe), and the 16S reference gene (200 nM probe). Primer and probe sequences used to detect certain identified bacterial species (including the ten selected bacterial species described herein) and the reference gene are listed in Table 2.1. The ddPCR mixture consisted of the following: 10 μL of ddPCR probe premix (dUTP-free, Bio-Rad, catalog number 1863024), primers (900 nM), probe (concentration listed herein), 2 μL of DNA (diluted to 5 ng / μL in reactions 1 and 2, and to 0.05 ng / μL in reaction 3), and nuclease-free water (to 20 μL). Then, 20 μL of the ddPCR mixture and 70 μL of probe were loaded into a DG8™ box (Bio-Rad, catalog number 1864008) using droplet generation oil (Bio-Rad, catalog number 1863005). A DG8™ gasket (Bio-Rad, catalog number 1863009) was hooked onto the box holder. Droplets for each sample were generated using a QX200 droplet generator (Bio-Rad) and then transferred to a ddPCR™ 96-well plate (Bio-Rad, catalog number 12001925). After sealing the plate with aluminum foil (Bio-Rad Laboratories, catalog number 1814040) and sealing it with PX1 PCR plate sealant (Bio-Rad Laboratories), the 96-well plate was run on a Bio-Rad Laboratories T100 PCR system. The PCR program was as follows: (1) pre-denaturation at 95°C for 10 min; (2) denaturation at 94°C for 30 s, followed by annealing and extension at 59°C for 1 min, for 40 cycles; (3) enzyme inactivation at 98°C for 10 min. After PCR, the fluorescence of the droplets was detected in a QX200 droplet reader (Bio-Rad Laboratories), and the data were analyzed using QuantaSoft™ Analysis Pro (v 1.0.596). The concentration of each bacterial species was determined and then normalized to the concentration of a 16S reference gene (normalized abundance of target = target concentration / 16S concentration). Table 2.1: Nucleotide sequences of primers and probes targeting bacterial species. Table 2.1 lists the nucleic acid sequences of the PCR products amplified by the primers targeting bacterial species.

[0609] Table 2.1 lists the nucleic acid sequences of the PCR products (“amplified fragments”) amplified by the primers for the selected bacterial species, as well as the nucleic acid sequences of the genes containing these amplified fragments (“marker genes”).

[0610] Bileophila warwickii

[0611] Amplified fragment: Whole genome shotgun sequence of region 202065 to 202184 of *Biliobacterium warfarinii* ATCC 49260 T370DRAFT_scaffold00002.2_C (GenBank: JNJP01000002.1).

[0612] AGGAAGCCGTCGAAAAGTACGCCGACAAGATGAT ACTCATGGGCACCATCGATCCCGGTTGGTTCATGACCGCGACTCCTGAAACCCTGAAGGCGCGCGTGGAAGAAGAACTTGCCGTCT (SEQ ID NO: 136)

[0613] Biomarker gene: T370_RS0102035 (1044 nt)

[0614] Gene product: Uroporphyrinogen decarboxylase family protein

[0615] Nucleic acid sequence:

[0616]

[0617] Clostridium perfringens

[0618] Amplified fragment: Region 59277 to 59356 of *Clostridium perfringens* DSM 753 3c, whole genome shotgun sequence (GenBank: NOXF01000003.1)

[0619] Nucleic acid sequence:

[0620] GGACTGAGCTTTCCAACCTTCACAGCATTCCCAC CGATTGCCCGCAGCGCGATGAACGTCTAGGCTGGCTGAATGACATG (SEQ ID NO: 138)

[0621] Biomarker gene: CH238_05410 (2664 nt)

[0622] Nucleic acid sequence:

[0623]

[0624] Clostridium sacchariformis

[0625] Amplified fragment: Regions 75654 to 75747 of the genome of Clostridium sacchariformis strain 2789STDY5834923, contig: SCcontig000005, whole genome shotgun sequence (GenBank: CZBB01000005.1).

[0626] Nucleic acid sequence:

[0627] TGGTTGCACGAATACCCATTTTGTCAATGCAAAC GGTCTGCCAGATCCGAACCACTACACCAGTGCACACGATATGGCACTCATTATGCAGGAG (SEQ ID NO: 140)

[0628] Biomarker gene: dacB_3 (1308 nt)

[0629] Gene product: D-alanyl-D-alanine carboxypeptidase dacB precursor

[0630] Nucleic acid sequence:

[0631]

[0632] Formic acid gemi

[0633] Amplified fragment: Regions 1148 to 1265 of the genome assembly of *G. formic acid* strain ATCC 27749, contiguous group: EI45DRAFT_scaffold00074.74, whole genome shotgun sequence (GenBank: FUYF01000074.1).

[0634] Nucleic acid sequence:

[0635] GGAGAAGAGGGAAAGCAGTCACTTTCTCGAATT GATGCCGCATCTCTGGATGGAAGCAACCGTAAAACACTTATTACATTTAACAATGGGGAACGTCTTGGCGATGTGTTTTGTACCA (SEQ ID NO: 142)

[0636] Biomarker gene: SAMN02745178_02911 (1362 nt)

[0637] Nucleic acid sequence:

[0638]

[0639] Visceral Odor Bacteria

[0640] Amplified fragment: Regions 2137841 to 2137944 of the genome assembly of Bacteroides viscera strain NCTC10825, chromosome: 1 (GenBank: LT906459.1)

[0641] Nucleic acid sequence:

[0642] GACAATCAACTCGACGGGAAAGCCTGGTTCGATG CTACAGGCCAATGGTATATGACCGAAACCGAGTTGACCCACACCAGCCAACTGCCTGAGGACGTACAAAA (SEQ ID NO: 144)

[0643] Biomarker gene: SAMEA44545918_01834 (834 nt)

[0644] Gene product: putative lipoprotein

[0645] Nucleic acid sequence:

[0646] ATGAACTTATACGCATTATTTTTCGGATTAGCCCTA TTCGGCTTTTGCAGTTGCGACAACGACGATTCGGTAACCCCTGATCCTCAAGCCCGCCAAGCTTTTGAAAGTAAATATCCCAACGCCACACAAGTGGAATGGGAAAAGAAAAACAATTACCTGATAGCAGAATTCATCGACAATCAACTCGACGGGAAAGCCTGGTTCGATGCTACAGGCCAATGGTATATGACCGAAACCGAGTTGACCCACACCAGCCAACTGCCTGAGGACGTACAAAAAGCTTTGGCAAACAGTGAATATGCACAATGGTATATCGACGATATCGATCGTTTGGAGCGTAACGAAACCGAAACGATTTATGTCATCGAAGTGAAAAAAGACAAACAAGAATACGATCTCTATTATTCAGCCGACGGAATTTTGGTCAAAGTTCAGGCCGACCTAACCGATGACGATTACGAAAACTATCTTCCCGACGCCTCGGAATTACCGGCTTCGCTCCAGCAATTCATCAATGACAAATATCCGACTTGCCGTATCATCGAAACTGAAACCGAACATAACCGTACAGAGGTAGATATCATACACGAAAACCGCAGCAAAGAAGTGATATTCGATGCCTCCGGCAATTGGCTGAATACCCATTACGATGTCCGCCAAAACGAAGTTGAAACCGCCGTAATGAATGCCTTGAAAACTTCGGCTTATGCCGATTATATCATCGACGATATTGAACGTTACGAAACGCCGGATCAAACCTATTACCTGTTCGAACTCGAAAAAGGAGCCAAAGAAATCGAAATAAAAATAGATCTGTCAGGGAAACTGCTTTAA (SEQ ID NO: 145)

[0647] Ruminococcus twistis

[0648] Amplified fragment: region 2028515 to 2028624 in the draft genome of Ruminococcus torques L2-14 (GenBank: FP929055.1)

[0649] Nucleic acid sequence:

[0650] CATATGAGGCTCTGCTGCTTTTAGGAAAGACGAC AGATACGCAGGACATTTCCGGAGAGGTTCTGGAGGAGAGGGATCCGGGCGATCTGACGGAAGAGGAAGTACGTTCC (SEQ ID NO: 146)

[0651] Biomarker gene: RTO_19540 (927 nt)

[0652] Gene product: tRNA pseudouridine synthase B

[0653] Nucleic acid sequence:

[0654] ATGATAAACGGAATAGTTAATATTTACAAAGAAA(SEQ ID NO: 147)

[0655] Eubacterium species CAG: 274

[0656] Amplified fragment: Region 7396 to 7500 of Eubacterium species CAG:274, WGS project CBEX01000000 data, contigs, whole genome shotgun sequence (GenBank: CBEX010000022.1).

[0657] Nucleic acid sequence:

[0658] TCATTTGTTTGGGTGCTGTCTTAGGTCTTGTGTTG GTTACAAAAATAGCCAAGGCTGAGGGGCAGAACCCTGATACATATGTGGACTTCGCTATATGGGCTTTGA (SEQ ID NO: 71)

[0659] Biomarker gene: BN582_00070 (861 nt)

[0660] Gene product: prolipoprotein diacylglycerol transferase

[0661] Nucleic acid sequence: identical to the nucleic acid sequence of the biomarker gene BN582_00070 as described in Example 1.

[0662] Lawsonia sucrase

[0663] Amplified fragment: Regions 627394 to 627504 of chromosome OA10 of *Lactobacillus suis* strain, complete genome (GenBank: CP091870.1).

[0664] Nucleic acid sequence:

[0665] TTGACGCCTTGACCATATCCACTTTGAAGCACCTA CTGCGCCAGCTTCCGCTGCGTACGGACATTACCATGGAGGAAGTGGTGGACCTATTCCGGCAGTTTCAGGACTTTA (SEQ ID NO: 67)

[0666] Biomarker gene: L9O85_02835 (633nt)

[0667] Gene products: TetR / AcrR family transcription regulators

[0668] Nucleic acid sequence: identical to the nucleic acid sequence of the biomarker gene L9O85_02835 as described in Example 1.

[0669] CAG: 241 of the genus *Vibrio*

[0670] Amplified fragment: Regions 1167 to 1283 of CAG:241 of the genus *Oscillatoria*, WGS project CBDI01000000 data, contigs, whole genome shotgun sequence (GenBank: CBDI010000003.1).

[0671] Nucleic acid sequence:

[0672] ACAGCCTCTGGTCATCCTTATACTCCACGTAGGCA TTGGCGTCATCCGCGAAAAAGTAATTCGCGTAGTCTTTTTTGTAACTATTAGAAAATGCCACGGGGTATCTGATATTCGCGG (SEQ ID NO: 68)

[0673] Biomarker gene: BN557_01301 (1506 nt)

[0674] Nucleic acid sequence: identical to the nucleic acid sequence of the biomarker gene BN557_01301 as described in Example 1.

[0675] Hidden glycolytic bacteria

[0676] Amplified fragment: Regions 1205228 to 1205347 of the DNA of the estrol-producing subspecies of Adlerella cloisterrum JCM 14811, complete genome (GenBank: AP024470.1).

[0677] Nucleic acid sequence:

[0678] GCCATTTCTCGCCGTTTAAGATATAGTTGCCCTCG GCGTCCTTGGTGGCCGTGGTGTGCACGTCGGCGCCGGTGCCGCCGGTCTGCTCGGTGAGCGCGAAGCAGGTGAAGATGGTCTTGT (SEQ ID NO: 149)

[0679] Biomarker gene: ADLECEL_09850 (1215 nt)

[0680] Gene product: Acyl-CoA dehydrogenase

[0681] Nucleic acid sequence:

[0682]

[0683] Toxic butyric acid bacteria

[0684] Amplified fragment: Region 4511686 to 4511794 of chromosome 23226 of *Vibrio butyricum* strain DSM, complete genome (GenBank: CP102269.1)

[0685] Nucleic acid sequence:

[0686] GGACGGAATGTAAGGGAGAATTTTGAGAAGACGA AAGAGGTGGTGGATGCGCTTGCGGCTTCTAAATTGGCTCAACTGGATACTCTAAAGGGGGTAACGGAGCTTTTTA (SEQ ID NO: 151)

[0687] Biomarker gene: NQ494_18680 (1347 nt)

[0688] Gene product: TlpA family protein disulfide reductase

[0689] Nucleic acid sequence:

[0690]

[0691] Clostridium species CAG: 58

[0692] Amplified fragment: Regions 14734 to 14849 of Clostridium species CAG:58, WGS project CBFK01000000 data, contigs, whole genome shotgun sequence (GenBank: CBFK010000011.1).

[0693] Nucleic acid sequence:

[0694] ACGCAGAATCTGTATCGCTTTCAGAAAAGCCGTGG AACCGCATCGATGGGGATACGTTCGGGAGCTTTTCCGATGTGGAGATCGGGATCCGGGCAGCAGGAGAAAGAAGCCTCTGG (SEQ ID NO: 153)

[0695] Biomarker gene: BN719_00886 (483 nt)

[0696] Nucleic acid sequence:

[0697] ATGAAGGGGGACCAGGGGATTTACGCTGGTGGAG CTGATGGCGGCTGCCATGCTGGCAGGAGCGGCGCTTTTGACCGCCCTTCCCCTGATGCTCATGGCCCAGCGGACTGCCCGCAGGCTGTCCGTGAAACAGCAGCAGCCATGAGAGGGGACGCGGTCTTTGAGCTGGCAGCCGGGGAGCTCAGATACGCAGAATCTGTATCGCTTTCAGAAAAGCCGTGGAACCGCATCGATGGGGATACGTTCGGGAGCTTTTCCGA TGTGGAGATCGGGATCCGGGCAGCAGGAGAAAGAAGCCTCTGGCTGACGGTACGGGTAAACCAGGGAGAAGAGCTTCTATATGAGAGGACTGGGAGGGTGGAGCTTCTGAACGTTCCTTTTTATGGAACTGGAAGGCTCGAAGGCAGCGGAAGGACCGGGCAGTCTGCCGCGGAGGGAGCTTCCTGTCCGGCGGTCTGGTACATAGACGGAGGAGATTCATGA (SEQ ID NO: 154)

[0698] fecal Collins

[0699] Amplified fragment: Region 39110 to 39203 of chromosome 13279 of *Collins faecalis* strain DSM, complete genome (GenBank: CP102276.1)

[0700] Nucleic acid sequence:

[0701] GAGAACCACCTGAACCTCACCAACACGAGCCGT CTCGAGCCCGAGGACATCTTGGACGAGGAGTTCTTCGACGACGACCTCGACGATATGGATG (SEQ ID NO: 155)

[0702] Biomarker gene: NQ498_00160 (1653 nt)

[0703] Gene product: C69 family dipeptidase

[0704] Nucleic acid sequence:

[0705]

[0706] fecal koala bacteria

[0707] Amplified fragment: Regions 2275359 to 2275458 of the DNA of *Koala faecalis* JCM 30894, complete genome (GenBank: AP019004.1)

[0708] Nucleic acid sequence:

[0709] GTTCAAGGACACACGGATCAAGGTGGCGCTGGTC ATGGACGAATTCGGCGGGCTGGCTGGTATGGTGACCCTGCGTGATATTATTGAGCATCTGGTCGGC (SEQ ID NO: 157)

[0710] Biomarker gene: corC_2 (1341 nt)

[0711] Gene product: magnesium and cobalt efflux protein CorC

[0712] Nucleic acid sequence:

[0713]

[0714] Broutella Hansenii

[0715] Amplified fragment: Region 333953 to 334071 of chromosome 20583 of *Brutella Hansenii*, complete genome (GenBank: CP022413.2).

[0716] Nucleic acid sequence:

[0717] ACGTGTAGAGTGGGCAAAAACAAAGGATGCTGA CGTATTTATCAGCTTTCATATCAATTCTTCGGTTTCCGGTGCAGCGCAGGGAGCAGAAGTATATTATCCGCAAGGAGAGGAAAATG (SEQ ID NO: 159)

[0718] Biomarker gene: CGC63_01640 (3174 nt)

[0719] Gene product: cell wall binding protein

[0720] Nucleic acid sequence:

[0721]

[0722] Clostridium spiralis (Thomasclavelia spiroformis)

[0723] Amplified fragment: Region 652607 to 652721 of chromosome 1552 of *Clavula helicum*, complete genome (GenBank: CP102275.1).

[0724] Nucleic acid sequence:

[0725] CAAAGCTATTGCAGTTGGACATGATTTAGGTCATG CTCCGTTGGTCATGGTGGAGAGCGTATAATTGATGAAATTGCAAAGGAATATGGATTACAATCGTTCTGGCATGAGCAA (SEQ ID NO: 161)

[0726] Biomarker gene: NQ543_02855 (1245 nt)

[0727] Gene product: Protein containing the HD domain

[0728] Nucleic acid sequence:

[0729]

[0730] Measles twin cocci

[0731] Amplified fragment: Regions 1033945 to 1034042 of the genome assembly of Genomic ...

[0732] Nucleic acid sequence:

[0733] TGGAACAGGAATGGAGCATCGACTCTAACCGTG CGCTAATGGGAGCGAACATGCAACGTCAAGCAGTGCCTTTAATGTTAACAGAGGCACCATTTGTTGGAACAGGAATGGAGCATC (SEQ ID NO: 163)

[0734] Biomarker gene: rpoB (3819 nt)

[0735] Gene product: DNA directed RNA polymerase subunit β

[0736] Nucleic acid sequence:

[0737]

[0738] Actinomycetes species oral taxonomy group 181

[0739] Amplified fragment: Regions 458 to 544 of strain 181 of the oral taxa of the genus Actinobacteria. F0379 A_sporaltaxon181-1.0_Cont73.5, whole genome shotgun sequence (GenBank: AMEW01000023.1)

[0740] Nucleic acid sequence:

[0741] CGTGTTGGAGTGCTATTTGCGTCCGGATGCTGGTC CGCTCCTGAGTCAGTGGGTTGATCTTGAGTGGACGCCTGAGGAGGATGAATA (SEQ ID NO: 74)

[0742] Biomarker gene: HMPREF9061_01114 (545 nt)

[0743] Nucleic acid sequence: identical to the nucleic acid sequence of the biomarker gene HMPREF9061_01114 as described in Example 1. Example 3: Performance determines that the gut bacterial profile differs between UC patients and normal individuals.

[0744] The gut microbiota differs between UC patients and healthy individuals.

[0745] Now for reference Figure 13A The graph shows the top bacterial species identified in the study as being associated with ulcerative colitis (UC). The lollipop plot in the left subplot shows the coefficient of each species with disease calculated using MaAsLin2, adjusted for age and sex. The boxes in the middle subplot represent phyla for each species. The bar chart in the right subplot shows the proportion of each species present in both the UC and healthy groups.

[0746] like Figure 13AAs shown, MaAsLin2 analysis revealed a negative correlation between 48 bacterial species and UC, namely UC-depleted species, including *Colinsella aerogenes*, *Clostridium tenuifolium*, *Ruminococcus tortifolius*, *Bacillus crypticarum*, *Gemacobacter formic acid*, *Clostridium sacchariformis*, *Alternaria putrefactiveis*, *Dolphobicella longiformis*, *Ruminococcus brucelli*, *Ostomyces viscerans*, *Trichophyton floccosum*, *Codonopsis pilosula*, *Codonopsis spp.*, *Codonopsis foetida*, *Adelaide cloacae*, *Roseidonella inulinophila*, *Brutella ovalis*, *Eubacterium rectum*, *Dolphobicella formic acid*, *Alternaria sarcopenia*, *Bacteroides foetida*, *Bacteroides foetida*, and *Oscillatoria* species (CAG). 241. Akkermansia myxophila, Alternaria fibrillosa, Eubacterium hominis, Species 57-20, Rosella humanis, Collagen foetida, Collins foetida, Bifidobacterium adolescentis, Rosella enterica, Toxic Butyrosporum tobira, Alternaria unobservatum, Species CAG: 38, Aerobic bacillus hardrovirus, Bacteroides fecalis, Species CAG: 242, Lawsonia sacchariformis, Species CAG: 274, Fecal cocci, Bacteroides masei, Rosella fecalis, Species CAG: 58, Lactobacillus lactis, Bacteroides vulgaris, Bileophilus warwickii, Firmicutes CAG: 145, Eubacterium clematis and Enterobacteriaceae cecum.

[0747] Fifteen bacterial species were found to be positively correlated with UC, namely UC-enriched species, namely Geminid measles, Veillonella infantis, Haemophilus species HMSC71H05, Actinomyces species oral taxa 181, Breutzia aerogenes, Lactobacillus mucosa, Enterococcus avium, Veillonella atypical, Eubacterium gingivalis, Breutzia hensenii, Roche myxobolus, Clostridium spiralis, Tizel bacillus, Veillonella septum, and Bacteroides fragilis.

[0748] Of the 48 UC-depleted bacterial species described in this paper, 14 were identified as potential bacterial biomarkers for UC diagnosis and were found to be negatively correlated with UC. These UC-depleted species are: *Koalacus feces*, *Bacillus crypticarum*, *Colinus feces*, *Oscillatoria* species CAG: 241, *Lawsonia crypticarum*, *Toxococcus butyricum*, *Clostridium* species CAG: 58, *Eubacterium* species CAG: 274, *Bacillus warfarinii*, *Clostridium tenuifolium*, *Clostridium sacchariformis*, *Gemmelia formica*, *Ostomyces viscerans*, and *Ruminococcus tortifolia*. The mean relative abundance (%) of UC-depleted species in normal individuals is shown in Table 2.2.

[0749] Of the 15 UC-enriched bacterial species described in this article, four were also identified as potential bacterial markers for UC diagnosis and were found to be positively correlated with UC. These are the UC-enriched species: Actinomycetes oral taxa 181, Clostridium spiralis, Gemstoneella mesenteriae, and Broutella hensenii. The mean relative abundance (%) of UC-enriched species in normal individuals is shown in Table 2.3.

[0750] like Figure 13B The diagram shows the associations between UC, sex, age, and the relative abundance of 14 UC-depleted bacterial species and 4 UC-enriched species, calculated using the generalized linear model MaAsLin2. Significant correlations (FDR < 0.05) are indicated by a plus sign for positive correlations and by a minus sign for negative correlations. The false discovery rate (FDR) was calculated using Bengemini-Hochberg correction. These bacteria can be used to determine the risk or diagnose ulcerative colitis in subjects.

[0751] Therefore, the bacteria listed in Tables 2.2 and 2.3 can be used in different combinations to establish an assessment model to determine the risk or diagnosis of ulcerative colitis in subjects, and whether microbiome recovery therapy or supplementation is needed. Relative abundance can be determined by using a set of primers for qPCR or ddPCR or by metagenomic sequencing to determine the risk or diagnosis of ulcerative colitis in subjects.

[0752] The bacteria listed in Table 2.2 can be used to relieve UC symptoms in UC patients. Table 2.2: Bacterial species enriched in normal individuals compared to UC patients (UC depleted species). Table 2.3: Bacterial species enriched in UC patients compared to normal individuals (UC-enriched species)

[0753] Machine learning models based on metagenomic data, using single bacterial biomarkers or combinations of different bacterial biomarkers. Performance.

[0754] The performance of machine learning models for UC diagnosis was evaluated using single bacterial biomarkers or combinations of different bacterial biomarkers. Model performance for single bacterial biomarkers ranged from 0.344 to 0.799 (numbers 1-18 in Table 2.4). To enhance the model's disease diagnostic capability, different combinations of bacterial biomarkers were tested, with AUCs ranging from 0.7 to 0.903 in the discovery cohort (numbers 19-132 in Table 2.4). A combination of ten bacterial biomarkers achieved the best diagnostic performance, with an AUC of 0.903 (number 19 in Table 2.4). Table 2.4: Machine learning models based on different combinations of bacterial biomarkers using metagenomic data for UC risk Predicted performance.

[0755] Performance of machine learning models based on metagenomic data and ten selected bacterial biomarkers.

[0756] Now for reference Figures 14A-14BThe table shows 18 bacterial species markers identified by metagenomic sequencing (as listed in Tables 2.2 and 2.3) in a cohort found in Hong Kong, China. Figure 14A ) and Hong Kong, China verification queue ( Figure 14B The relative abundance (%) of UC patients and normal individuals.

[0757] To obtain models with optimal performance across different populations, various combinations of bacterial biomarkers were evaluated. Ultimately, ten selected bacterial species biomarkers (also referred to as "biomarkers") were used in the machine learning model, including *Biliobacterium warfarinii*, *Clostridium tenuifolium*, *Clostridium sacchariformis*, *Gemmelella formica*, *Ostomyces viscerans*, *Ruminococcus tortifolia*, *Actinomyces* species oral taxa 181, *Clostridium spiralis*, *Genitia measles*, and *Brutella hensenii* (as listed in Table 2.5). Figures 14A-14B As shown, the relative abundance of these 10 biomarkers is consistent in the Hong Kong discovery cohort and the Hong Kong validation cohort.

[0758] Now for reference Figure 15A The figure shows the receiver operating characteristic (ROC) curve and area under the curve (AUC) of the machine learning model. AUCs were calculated using a random forest model with 10 biomarkers identified through metagenomic sequencing in the training set, test set, and Hong Kong validation cohort.

[0759] The final model using these 10 markers found a queue in Hong Kong, China (e.g.) Figure 15A The area under the curve (AUC) of the receiver operating characteristic (ROC) curve in the test set (labeled as "test set") is 0.9025 (95% CI: 0.8431–0.9618; 88.06% sensitivity, 80.95% specificity, 0.564 optimized threshold). In the Hong Kong validation cohort, these 10 biomarkers (labeled as "validation cohort")... Figure 15A The AUC (shown in the figure) is 0.8075.

[0760] Now for reference Figure 15B The study shows the receiver operating characteristic (ROC) curves and area under the curve (AUC) values ​​of a machine learning model (random forest model) using 10 biomarkers identified through metagenomic sequencing on three public datasets including subjects from the United States, the Netherlands, and China. The AUCs of the diagnostic model in the US, Netherlands, and China were 0.8549, 0.8706, and 0.8213, respectively, on the three public datasets.

[0761] The results showed that the machine learning model using 10 biomarkers was accurate in predicting the risk of ulcerative colitis in the subjects. Table 2.5: Bacterial species included in the machine learning model for UC risk prediction

[0762] Determining the risk of ulcerative colitis using different combinations of bacterial biomarkers.

[0763] To determine the risk of ulcerative colitis or diagnose Crohn's disease in subjects, the following steps are performed: (1) a training dataset is obtained by determining the relative abundance of species selected from Table 2.2 or Table 2.3 in normal individuals and a cohort of patients with ulcerative colitis; (2) the relative abundance of these species is determined in the subjects whose risk of ulcerative colitis is to be determined; (3) the relative abundance of these species in the subjects is compared with the training data using a random forest model; and (4) a decision tree is generated from the training data using a random forest. The relative abundance is run along the decision tree and a risk score is generated. If more than 50% of the trees in the model consider the subject to have ulcerative colitis, the subject being tested is considered to have an increased risk of ulcerative colitis. If less than 50% of the trees in the model consider the subject to be a normal individual, the subject being tested is considered not to have an increased risk of ulcerative colitis.

[0764] Bacterial species identified by droplet digital PCR (ddPCR)

[0765] Now for reference Figures 16A-16C Using the specific primers and probes listed in Table 2.1, the levels of 6 bacterial species negatively correlated with UC and 4 bacterial species positively correlated with UC were determined by ddPCR in the subgroups of the discovery cohort (UC = 205, control = 84) (Table 2.5).

[0766] The use of ddPCR data to develop machine learning models based on single bacterial biomarkers or combinations of different bacterial biomarkers. able.

[0767] The performance of machine learning models for UC diagnosis was evaluated using single bacterial biomarkers or combinations of different bacterial biomarkers. Model performance for single bacterial biomarkers ranged from 0.403 to 0.789 (numbers 1-10 in Table 2.6). To enhance the model's diagnostic capabilities, different combinations of bacterial biomarkers were tested, with AUCs ranging from 0.614 to 0.881 (numbers 11-136 in Table 2.6). A combination of ten selected bacterial biomarkers achieved the best diagnostic performance, with an AUC of 0.881 (number 11 in Table 2.6). Table 2.6: Machine learning models based on different combinations of bacterial biomarkers using ddPCR data for UC risk prediction Performance.

[0768] Performance of machine learning models using ddPCR data

[0769] Ten selected bacterial species biomarkers were used in the machine learning model, including *Biliobacterium warfarinii*, *Clostridium tenuifolium*, *Clostridium sacchariformis*, *Gemmelia formica*, *Ostomyces viscerans*, *Ruminococcus tortifolia*, *Actinomyces* species oral taxa 181, *Clostridium spiralis*, *Gimeria measles*, and *Brutella hensenii* (Table 2.5). A subgroup of the cohort (UC = 205, control = 84) was randomly assigned to a training set (UC = 165, control = 62) and a test set (UC = 40, control = 22).

[0770] Figure 17 The receiver operating characteristic (ROC) curves and area under the curve (AUC) of the machine learning model are shown. The AUC of the random forest model using 10 biomarkers identified by droplet digital PCR is also shown. The final model using these 10 biomarkers achieved an AUC of 0.9521 on the training set and 0.8807 on the test set (95% CI: 0.7885–0.9729; 85.00% sensitivity, 81.82% specificity, optimized threshold of 0.658).

[0771] Biomarkers of depleted bacterial species in UC can help deplete beneficial pathways in UC patients.

[0772] Figures 18A-18D Stacked bar charts compared the relative abundance (%) of different bacterial species (i.e., Actinomyces species oral taxa 181, *Biliobacterium wartii*, *Broutella hensenopyrum*, *Clostridium velutipes*, *Clostridium spiralis*, *Clostridium sacchariformis*, *Twincocephalus mesenteriae*, *Gemmelella formica*, *Ostomyces viscerans*, *Ruminococcus tortifolius*, and other bacterial species) in the following functional pathways between healthy controls and UC patients: L-arginine biosynthesis I (via L-ornithine), L-valine biosynthesis, L-ornithine biosynthesis I, and de novo adenosine ribonucleotide biosynthesis. Each stacked bar chart represents the contribution of bacterial species biomarkers and other bacteria to each functional pathway.

[0773] like Figures 18A-18DAs shown, the abundance of the following functional pathways is reduced in UC patients: amino acid biosynthesis (L-arginine, L-ornithine, L-valine biosynthesis) and nucleoside and nucleotide biosynthesis (adenosylribonucleotide de novo biosynthesis). These pathways are abundant in healthy controls. The higher abundance of these pathways in healthy controls is mainly contributed by UC-depleted species, such as *Clostridium tenuifolium*, *Clostridium sacchariformis*, *Gemmelella formica*, and *Ruminococcus tortifolia*. Therefore, providing UC patients with effective amounts of *Clostridium tenuifolium*, *Clostridium sacchariformis*, and / or *Clostridium sacchariformis* can increase the abundance of these functional pathways, which is beneficial for the treatment of UC. Example 4: Predicting UC risk in test subjects

[0774] In this example, to determine the risk score of ulcerative colitis in subjects, the following steps were performed: (1) The relative abundance of 10 selected bacterial species biomarkers as shown in Table 2.5 was measured in the subjects by metagenomic sequencing or ddPCR; (2) The relative abundance data were input into a pre-trained machine learning model (random forest model) to generate a disease risk score (disease probability) for the subjects; (3) If the risk score (disease probability) of the subject was higher than the optimized cutoff value of 0.564 for metagenomic data or 0.658 for ddPCR data, the subject was considered to have ulcerative colitis or be at risk of having ulcerative colitis. Conversely, if the risk score (disease probability) was not higher than the optimized cutoff value, the individual was considered not to have ulcerative colitis or be at risk of having ulcerative colitis.

[0775] Table 2.7 shows the predicted UC risk (disease probability) for 109 subjects (including known UC patients and controls, as shown in the “Original Group”) using a machine learning model based on metagenomic data. Table 2.8 shows the predicted UC risk (disease probability) for 62 subjects (including known UC patients and controls, as shown in the “Original Group”) using a machine learning model based on ddPCR data.

[0776] The “Original Group” column indicates the condition of each subject as assessed by a healthcare professional through other means (e.g., colonoscopy), where “UC” refers to subjects diagnosed with “ulcerative colitis”, and “Control” refers to subjects without ulcerative colitis. After analyzing samples from subjects according to the disclosed method as described in the foregoing examples, a risk score (probability of disease or risk prediction) is calculated. If the risk score is higher than the optimized cutoff value, the subject is marked as “UC” in the “Test Results” column. Conversely, if the risk score is not higher than the optimized cutoff value, the subject is marked as “Control” in the “Test Results” column.

[0777] As shown in Table 2.7, in the test results using the machine learning model based on metagenomic data, 34 out of 42 subjects in the original "control" group were correctly identified as "controls," and 59 out of 67 subjects in the original "UC" group were correctly identified as "UC." As shown in Table 2.8, in the test results using the machine learning model based on ddPCR data, 18 out of 22 subjects in the original "control" group were correctly identified as "controls," and 34 out of 40 subjects in the original "UC" group were correctly identified as "UC."

[0778] The machine learning model based on metagenomic data achieved a sensitivity of 88.06% and a specificity of 80.95%. The machine learning model based on ddPCR data achieved a sensitivity of 85.00% and a specificity of 81.82%.

[0779] These results demonstrate that machine learning models based on metagenomic and ddPCR data are highly accurate in predicting the risk of UC in subjects. Table 2.7: UC risk scores (disease profile) of subjects when using machine learning models based on metagenomic data. Rate). Table 2.8: UC risk score (disease probability) of subjects when using machine learning models based on ddPCR data. Example 5: Risk prediction for UC patients with different disease activities Disease risk scores generated using a machine learning model employing ten selected UC bacterial biomarkers can reflect metabolic dysregulation in UC patients.

[0780] Each subject's metabolic dysfunction score was determined by calculating the median Bray-Curtis dissimilarity to the reference control group based on metabolic functional pathways. (Now referencing...) Figure 19 The figure shows a graph illustrating the correlation between functional ecodysregulation scores and disease probabilities generated by a machine learning model based on biomarkers of 10 selected UC bacterial species (according to Table 2.5). The correlation coefficients R and p-values ​​are given by Spearman correlation (…). (p < 0.05). Figure 19As shown, the disease risk score generated by a machine learning model using 10 selected UC bacterial biomarkers was positively correlated with the dysfunction score. This indicates that the machine learning model can reflect the metabolic dysfunction in UC patients. A diagnostic model based on ten selected bacterial biomarkers for UC can diagnose UC patients in remission.

[0781] UC is characterized by alternating periods of disease activity, including flare-ups and remissions. The accuracy and stability of 10 selected bacterial biomarkers (listed in Table 2.5) were evaluated in UC patients with varying disease activity. UC patients were categorized into those with active disease (Mayo > 2) and those with inactive disease (Mayo ≤ 2) using the Mayo Criterion Score. Figures 20A-20C The results showed that, compared with healthy controls, patients with inactive UC had a reduction in six UC-depleted bacterial markers: Clostridium truncatum, Clostridium sacchariformis, Gemma formic acid, Ruminococcus truncatum, Osmotherum visceratum, and Bacillus vortexe. Compared with healthy controls, patients with inactive UC had an increase in four UC-enriched bacterial markers: Geminidella mesenteriae, Broutella henseniella, Actinomyces species oral taxa 181, and Clostridium spiralis. Figure 20D The results showed that the disease risk scores generated by a machine learning model using 10 selected UC bacterial biomarkers did not differ significantly between UC patients in inactive (N = 110) and active (N = 12) states. Figure 20E The model achieved an AUC of 0.885 when classifying UC patients in remission against controls. This demonstrates that machine learning models can be used to determine the risk of ulcerative colitis in individuals, regardless of disease activity.

[0782] Of the 10 selected bacterial biomarkers, *Clostridium truncatum*, *Clostridium sacchariformis*, *Ostomyces viviparus*, and *Gemmelia formica* showed lower levels in active UC patients compared to inactive UC patients, while *Actinomyces* species *Oral Taxonomycetes 181* and *Clostridium spiralis* showed higher levels in active UC patients than in inactive UC patients. Therefore, these species can be used as bacterial biomarkers for monitoring disease activity in UC patients. For example, a set of biomarkers containing these species could be used to train machine learning models in cohorts of UC patients with both active and inactive disease, enabling their use as biomarkers for disease monitoring and prediction of disease onset. Example 6: Differentiating UC from other diseases A diagnostic model based on ten selected bacterial biomarkers for UC can differentiate UC from other gastrointestinal (GI) and non-GI diseases.

[0783] Given the common microbiota alterations across various diseases, validating the disease specificity of identified bacterial markers is crucial to ensure a low false-positive rate in IBD diagnoses. To this end, several non-IBD disease datasets were evaluated, comprising subjects with gastrointestinal diseases (n = 439) and non-gastrointestinal diseases (n = 291). The gastrointestinal diseases included colorectal cancer (CRC, n = 160), colorectal adenoma (CA, n = 162), and irritable bowel syndrome (diarrhea subtype, IBS-D, n = 117), while the non-gastrointestinal diseases included obesity (BMI > 28; n = 148) and cardiovascular disease (CVD; n = 143).

[0784] Now for reference Figures 21A-21B . Figure 21A The figure shows the ROC curves for classifying UC patients with other GI diseases, including CA (n = 162), IBS-D (n = 117), and CRC (n = 160). Figure 21A As shown, the UC diagnostic model / machine learning model (random forest model) using 10 selected bacterial biomarkers identified by metagenomic sequencing can distinguish UC subjects from patients with other GI diseases. The AUROC values ​​for colorectal adenoma (CA), irritable bowel syndrome (diarrhea subtype, IBS-D), and colorectal cancer (CRC) are 0.8893, 0.8649, and 0.7575, respectively. Figure 21B This is a comparison of disease probabilities (risk scores) generated for UC patients and patients with other GI diseases using a model based on biomarkers of 10 selected UC bacterial species. P-values ​​were calculated using the Wilcoxon rank-sum test (p < 0.05). p < 0.01; p < 0.001; p < 0.0001; ns, no significance). Figure 21B As shown, the disease probability (risk score) of patients with other GI diseases is significantly lower than that of UC patients, demonstrating that the machine learning model is accurate in distinguishing between UC patients and patients with other GI diseases. Figure 21C The figure shows the ROC curves that classify UC patients from those with non-IBD GI diseases, where the UC diagnostic model distinguishes UC subjects from those with other GI diseases with a mean AUROC of 0.8348.

[0785] Now for reference Figures 22A-22B . Figure 22A The figure shows the ROC curves for classifying UC patients with other non-GI diseases, including obesity (n = 148) and CVD (n = 143), and a comparison of their disease probabilities generated by a model based on biomarkers of 10 selected UC bacterial species. Figure 22A As shown, the same model can also distinguish UC patients from patients with other non-GI diseases, including obesity (n = 148) and cardiovascular disease (CVD) (n = 143), with AUROCs of 0.8436 and 0.8996 for obesity and CVD, respectively. Figure 22B The figure shows a comparison of disease probabilities (risk scores) generated in UC and other non-GI diseases using a model based on biomarkers of 10 selected UC bacterial species. P-values ​​were calculated using the Wilcoxon rank-sum test (p < 0.05). p < 0.01; p < 0.001; p < 0.0001; ns, no significance). Figure 22B As shown, the disease probability (risk score) of patients with other non-GI diseases is significantly lower than that of UC patients, demonstrating that the machine learning model is accurate in distinguishing between UC patients and patients with other non-GI diseases.

[0786] Figure 22C The figure shows the ROC curves for classifying UC patients against patients with all other non-IBD diseases (GI and non-GI diseases) (n = 730) using a machine learning model based on biomarkers of 10 selected bacterial species. Figure 22C As shown, the machine learning model can distinguish UC patients from patients with all other non-IBD diseases, with an AUROC of 0.8493.

[0787] The results of differentiating UC from GI diseases and other non-GI diseases from UC based on models using biomarkers of ten selected UC bacterial species are summarized in Table 2.9. Table 2.9: Performance of UC diagnostic models in distinguishing UC from other gastrointestinal (GI) and non-GI diseases

[0788] These results demonstrate that the diagnostic model encompassing 10 selected UC-related biomarkers exhibits satisfactory performance in distinguishing UC from all other non-IBD diseases (i.e., GI diseases and / or other non-GI diseases).

[0789] Among ten selected bacterial markers, *Ruminococcus tortifolia* was depleted in UC patients compared to other diseases, while *Clostridium spiralis* was enriched in UC patients compared to other diseases. Example 7: Methods used to identify, diagnose, prevent, or treat the following risks in an individual: ulceration Ulcerative colitis (UC)

[0790] Now for reference Figure 23 It shows an example of method 200 for determining, diagnosing, preventing, or treating ulcerative colitis (UC) in an individual, the method involving the following steps:

[0791] Step 210: Determine the relative abundance of a set of bacterial species in a fecal sample from the individual, wherein the set of bacterial species includes one or more bacterial species selected from the group consisting of: *Collagenia fecalis*, *Bacillus crypticarum*, *Colinus fecalis*, *Oscillatoria* species CAG: 241, *Lawsonia crypticarum*, *Toxococcus butyricum*, *Clostridium* species CAG: 58, *Eubacterium* species CAG: 274, *Bacillus warwickii*, *Clostridium truncatum*, *Clostridium sacchariformis*, *Gemmelia formica*, *Ostomyces viscerans*, *Ruminococcus tortifolia*, *Actinomyces* species oral taxonomy group 181, *Clostridium spiralis*, *Gimidae muscarinica*, and *Broutella henselae*.

[0792] Step 220: Compare the relative abundance of each bacterial species in the bacterial species set from the individual with the relative abundance of each bacterial species in the bacterial species set in the reference dataset to generate a risk score;

[0793] Step 230: If the risk score is higher than the cutoff value, then the individual is determined to have an increased risk of developing UC or already has UC;

[0794] Step 240: If it is determined that the individual has an increased risk of developing UC or is already suffering from UC, the individual may optionally be treated. Example 3

[0795] Despite recent advances in understanding the association between the gut microbiome and inflammatory bowel disease (IBD), the role of microbiome biomarkers in IBD diagnosis remains underexplored. This study yielded a microbiome-based diagnostic test for IBD. Using metagenomic data from 5,979 fecal samples from individuals with and without IBD across different regions and ethnicities, alterations in the microbiome in IBD were identified, and ten and nine bacterial species, respectively, were selected to construct diagnostic models for ulcerative colitis (UC) and Crohn's disease (CD). These diagnostic models achieved an area under the curve (AUC) greater than 0.90 when distinguishing IBD patients from controls in the discovery cohort. In a cross-ethnic validation cohort comprising IBD patients (UC: n = 817, CD: n = 1,065) and non-IBD subjects (n = 3,600), the UC and CD diagnostic models maintained satisfactory performance with AUCs of 0.78 and 0.72, respectively. Further, multiplex droplet digital polymerase chain reactions targeting selected IBD-associated bacterial species were developed, and the m-ddPCR-based model demonstrated numerically superior performance compared to fecal calprotectin in distinguishing UC and CD from controls (UC: 0.74 vs. 0.61; CD: 0.78 vs. 0.56). In this example, universal IBD-associated bacterial species shared across cohorts were identified, and the potential applicability of a multibacterial biomarker set as a non-invasive tool for IBD diagnosis was demonstrated.

[0796] Altered gut microbiota composition and metabolic pathways have been observed in IBD patients. However, the role of microbiome biomarkers in the diagnosis of IBD remains underexplored.

[0797] In this example, a microbiome-based IBD diagnostic test was developed. A comprehensive analysis of metagenomic datasets was performed to assess the predictability of selected bacterial species for IBD diagnosis. A diagnostic model using disease-specific species was constructed, and a multibacterial biomarker set based on multiplex droplet digital PCR (m-ddPCR) for IBD diagnosis was developed, such as… Figure 24A As shown in the image.

[0798] Figures 24A-24CThe diagrams and figures illustrate an overview of the study workflow and a comparison of the fecal microbiomes in UC, CD patients, and controls. Some elements were created using BioRender.com. A total of 5979 samples were included in this study, including 1884 samples from the internal sequencing dataset and 4095 samples from the public dataset. The discovery cohort included 174 CD patients, 205 UC patients, and 118 controls. The independent IBD cohort included 139 UC patients, 190 CD patients, and 328 controls. The public dataset included 678 UC patients, 875 CD patients, and 1699 controls. The non-IBD cohort included 146 IBS patients, 230 CA patients, 372 CRC patients, 318 obese patients, 143 CVD patients, and 364 corresponding controls. Figure 24B This is a violin plot showing the Shannon index and observed species of the fecal microbiome in UC patients (N = 205), CD patients (N = 174), and controls (N = 118). Data are presented as box plots with the median (center line), 25th and 75th percentiles (box boundaries), and 5th and 95th percentiles (whiskers). P-values ​​were calculated using a two-sided Wilcoxon rank-sum test. Figure 24C Principal coordinate analysis (PCoA) plots showed the distinct microbial composition among the groups (174 CD patients, 205 UC patients, and 118 controls). Data were presented as box plots with the median (center line), 25th and 75th percentiles (box boundaries), and 5th and 95th percentiles (whiskers). β-diversity p-values ​​based on Bray-Curtis distance were calculated using PERMANOVA with 999 permutations (Df = 2, R² = 0.02219, F = 5.606, P = 0.001). Figure 24C' The multivariate analysis plot shows the amount of explained variance and the corresponding p-values ​​for species-level Bray-Curtis dissimilarity determined by PERMANOVA.

[0799] Figure 24D The stacked bar chart shows the relative abundance of the six most abundant phyla in UC patients (N = 205), CD patients (N = 174), and controls (N = 118). "Other" represents phyla not shown in the figure. CD: Crohn's disease; UC: ulcerative colitis; CA: colorectal adenoma; CRC: colorectal cancer; IBS: irritable bowel syndrome; CVD: cardiovascular disease. p < 0.05; p < 0.01; p < 0.001. Figure 24D'This is a comparison plot of the relative abundance of the six gates in UC patients, CD patients, and healthy controls. Box plots represent the minimum, Q1, median, Q3, and maximum values. P-values ​​were calculated using a two-sided Wilcoxon rank-sum test. p < 0.05, p < 0.01, p < 0.001, p < 0.0001, ns not significant. CD: Crohn's disease; UC: ulcerative colitis. result

[0800] Characterization of gut microbiome diversity and taxonomy in IBD

[0801] Fecal metagenomic data from 4,406 samples across 13 IBD cohorts in eight countries and regions were analyzed to identify and validate gut microbiota biomarkers for IBD diagnosis. Specifically, internal sequencing data from Hong Kong, China, were used as the discovery cohort (Table 3.1), and three additional independent internal cohorts from Hong Kong and Australia, as well as nine public datasets from the United States, the Netherlands, China, Spain, Denmark, and the United Kingdom, were included as validation cohorts (Table 3.2).

[0802] In the discovery cohort, a total of 1,175 taxa (3 kingdoms, 14 phyla, 25 classes, 40 orders, 85 families, 226 genera, and 788 species) were identified. At the species level, 674, 637, and 506 bacterial species were identified in UC, CD, and the control group, respectively. Compared with the control group, microbial diversity (median: UC 2.73, CD 2.71, control 3.08; P < 0.001) and richness (median: UC 82, CD 86.5, control 95; P < 0.001) were decreased in UC and CD patients, but there was no statistically significant difference between UC and CD. Figure 24B As shown in the figure. In principal coordinate analysis (PCoA) based on Bray-Curtis distance, the gut microbiota of UC and CD patients clustered separately from the gut microbiota of controls, and UC patients showed a greater distance from controls than CD patients, as shown in the figure. Figure 24C As shown in the figure. The presence of IBD explained 2.22% of the microbiome variation (P < 0.001), while age and sex contributed 0.28% (P = 0.107) and 0.29% (P = 0.073), respectively. Factors explaining microbiome variation are shown in the figure. Figure 24C'Significant differences were found in the gut microbiota at the phylum level between IBD patients and controls, characterized by a decrease in Firmicutes and an enrichment of Proteobacteria in IBD patients. Compared with UC and controls (both P < 0.001), CD patients had lower levels of Bacteroides (…). Figure 24D and Figure 24D' Compared with controls, patients with IBD were found to have reduced microbial diversity, and patients with CD and UC showed significant differences in their gut microbiome composition.

[0803] Identification of gut microbiome characteristics in UC and CD

[0804] Next, using a generalized linear model as implemented in MaAsLin2, bacterial species with differential abundance in UC and CD were identified after filtering out low-prevalence species (<10%) and adjusting for age and sex. Figure 25C-25H The figure shows the differential bacterial species and metabolic pathway dysbiosis in UC and CD patients compared to controls. Figure 13A and 1A The figures show the leading bacterial species associated with UC and CD, respectively. The lollipop plot in the left subplot shows the coefficient of each species with disease calculated using MaAsLin2, adjusted for age and sex. The boxes in the middle subplot represent the phyla for each species. The bar chart in the right subplot shows the proportion of each species present in UC, CD, and the control group. Figure 25C The figure shows the relative abundance of ten bacterial species biomarkers in the UC (N = 205) and control group (N = 118), and Figure 25D The figure shows the relative abundance of nine bacterial species biomarkers identified by metagenomics in CD (N = 174) and the control group (N = 118). Data are presented as box plots with the median (center line), 25th and 75th percentiles (box boundaries), and 5th and 95th percentiles (whiskers). P-values ​​were calculated using a two-tailed Wilcoxon rank-sum test. Figure 25E-25F The figure shows the performance of models with ten UC and nine CD bacterial species biomarkers, respectively, in classifying UC and CD patients in the discovery cohort against controls. The shaded area of ​​the ROC curve represents the 95% confidence interval of the AUC for the test set. Figure 25G-25HThe figure shows the Shapley Additive Explanation (SHAP) values ​​of bacterial species biomarkers in UC and CD for each sample. Each point represents the SHAP value of each biomarker for each sample. The distribution of points indicates the effect of each biomarker on the model output. Colors represent the relative abundance of biomarkers (yellow for high abundance, purple for low abundance). The results showed that 126 and 161 species, respectively, had differential abundance in UC and CD compared to the control (FDR < 0.25). Among these, 15 species, including *Bacteroides fragilis*, *Veillonella spp.*, *Taizei spp.*, *Clostridium spiralis*, *Rhodotorula spp.*, and *Brutella henselae*, were enriched in UC (FDR <= 0.1, coefficient >= 0.1), while 48 species, including *Collinsella oxyphylla*, *Clostridium tenuifolium*, *Ruminococcus truncatula*, *Bacillus spp.*, *Gemmelella formica*, and *Clostridium spp.*, were depleted in the intestines of UC patients (FDR < 0.05, coefficient <= -0.45). Figure 13A As shown in the diagram. In CD, 58 bacterial species were enriched or depleted compared to the control (FDR < 0.001 and |coefficient| >= 0.5). In particular, certain bacterial species with anti-inflammatory properties, including *Eubacterium hopterii*, *Brutella ovalis*, *Roseola inulin*, and *Eubacterium rectum*, were depleted in CD, such as... Figure 1A As shown in the figure. Furthermore, *E. coli* and some *Streptococcus* species were enriched in the intestines of patients with CD but not UC. In summary, these data highlight the presence of disease-specific bacterial species in both UC and CD.

[0805] Development of metagenomics-based diagnostic models for UC and CD diagnosis

[0806] Next, five-fold cross-validation was performed on all discriminative bacterial species to construct a diagnostic model. Stable classification performance was achieved in UC and CD by utilizing seven and eight bacterial features, respectively. Figure 25K and Figure 25L As shown. Figure 25K As shown, a total of 125 species features were used in the UC diagnostic model. The vertical dashed line in x = 7 represents the minimum number of features required to maintain the relative stability of the model (horizontal dashed line, AUC = 0.8937). Figure 25LAs shown, a total of 161 species features were used in the CD diagnostic model. The vertical dashed line in x = 8 represents the minimum number of features required to maintain the relative stability of the model (horizontal dashed line, AUC = 0.9096). After considering the functional characteristics of the selected bacteria and the optimal distribution of enriched and depleted bacterial species, ten bacterial species were selected as biomarkers for UC (4 enriched: *Germococcus mesenteriae*, *Broutella hensenii*, *Actinomyces* species oral taxa 181, *Clostridium spiralis*; 6 depleted: *Clostridium tenuifolium*, *Clostridium sacchariformis*, *Gemmelella formica*, *Ruminococcus tortifolia*, *Ostomyces viscerans*, *Biotrophozoites warwickii*) and nine bacterial species were selected as biomarkers for CD (3 enriched: *Bacteroides fragilis*, *Escherichia coli*, *Actinomyces* species oral taxa 181; 6 depleted: *Roseobacterium inulinum*, *Broutella ovalis*, *Lawsonia solani*, *Roseobacter esculenta*, *Dolphinium formateriate*, *Eubacterium* species CAG: 274). Figure 25C and Figure 25D Among these, *Actinomyces* species oral taxa 181 and *Clostridium spiralis* were reported enriched in UC for the first time, while *Actinomyces* species oral taxa 181, *Lawsonia sucrase*, and *Eubacterium* species CAG: 274 are novel bacterial species associated with CD. Using a random forest algorithm, ten selected bacterial species were used to distinguish UC patients from controls, with an AUC of 0.95 (95% CI: 0.92–0.98) in the training set and 0.90 (95% CI: 0.84–0.96; sensitivity 88.06%, specificity 80.95%) in the test set. *Clostridium sucrase*, *Clostridium tenuifolium*, and *G. formic acid* were the top three key identifying bacterial species in the UC model. Figure 25E and Figure 25G In CD, nine bacterial species showed AUCs of 0.95 (95% CI: 0.92–0.98) and 0.94 (95% CI: 0.89–0.98; sensitivity 88.33%, specificity 89.47%) in the training and test sets, respectively, distinguishing CD from controls. *Brutella ovalis*, *Lawsonia solani*, *Roseidonia inulinica* (deleted in CD), and *Actinomyces* species oral taxa 181 and *Escherichia coli* (enriched in CD) were the top-ranking bacteria in our model. Figure 25F and Figure 25H ).

[0807] Functional and metabolic pathways of bacterial species in UC and CD

[0808] In addition to alterations in microbial composition and taxonomy, metabolic dysfunction was also found in UC and CD. Using MaAsLin2 comparative analysis, 545 metabolic pathways were identified in UC and CD compared to controls, consisting of 244 and 315 differentially expressed pathways, respectively (FDR < 0.05).

[0809] Compared with controls, pathways involved in amine and polyamine degradation, as well as fatty acid and lipid biosynthesis, were significantly enriched in patients with UC and CD. Biomarkers of bacterial species enriched in UC and CD were positively correlated with disease-enriched metabolic pathways, while biomarkers of bacterial species depleted in UC and CD showed a negative correlation. Figures 26A-26D ). Figures 26A-26D The figure shows the differentially functional pathways between UC / CD patients and healthy controls and their correlation with bacterial species biomarkers. Differentially functional pathways were identified using MaAsLin2 (adjusted for age and sex). Correlation coefficients and p-values ​​between ten UC or nine CD bacterial species biomarkers and the differentially functional pathways are given by Spearman correlation. p < 0.05; p < 0.01; p < 0.001.

[0810] Results from stratified analysis showed that pathways involved in amino acid biosynthesis (L-arginine, L-ornithine, and L-valine biosynthesis) were primarily contributed by depleted bacterial species in UC patients, including *Clostridium tenuifolium*, *Clostridium sacchariformis*, *Gemmelella formica*, and *Ruminococcus tortifolia*. The depletion of these key bacterial species in UC was also associated with a significant decrease in the abundance of their respective functional pathways (for UC, ...). Figures 18A-18D Similarly, bacterial species, including *Brutella ovalis*, *Roseidon*, *Dolphobicella formica*, *CAG: 274*, and *Roseidon*, also contributed to the following pathways in controls: amino acid biosynthesis (L-tryptophan biosynthesis), carbohydrate degradation (starch degradation), cofactors, carriers, and vitamin biosynthesis (thiamine pyrophosphate formation from pyridine thiamine and hydroxy thiamine). In CD, the major contributors to these pathways have shifted, characterized by the dominance of *Escherichia coli* rather than a diverse bacterial spectrum (for CD, Figures 6A-6D Changes in bacterial metabolic pathways suggest that altered microbiome-mediated metabolic capacity may be crucial for the development of IBD.

[0811] To investigate the role of bacterial biomarkers in mediating metabolic function, a dysfunction score was developed for each subject using metabolic pathway profiling to calculate the median Bray-Curtis dissimilarity to controls. The distribution and levels of dysfunction scores differed between patients with UC and CD and controls, indicating functional changes in IBD. Figure 26E-26F The disease probability determined by the diagnostic model was positively correlated with dysfunction scores in UC and CD, implying that bacterial biomarkers could reflect metabolic dysregulation in IBD patients. Figure 19 and Figure 7 ).

[0812] Diagnostic performance of multibacterial biomarkers in host inflammation

[0813] The accuracy and stability of the multibacterial biomarker group in IBD with different disease activities were evaluated. Patients were divided into active disease (UC: Mayo > 2; CD: CDAI > 150) and inactive disease (UC: Mayo ≤ 2; CD: CDAI ≤ 150) using the Mayo Criterion Score and Crohn's Disease Activity Index (CDAI). Compared with controls, the abundance of CD and UC-related bacterial species differed in inactive IBD patients (6 species decreased and 4 increased in UC; 6 species decreased and 3 increased in CD). Furthermore, the levels of some bacterial species varied with disease activity. Figures 27A-27F The figures show the relative abundance of bacterial species biomarkers in healthy controls and patients in inactive and active states. The shaded areas of the ROC curves represent the 95% confidence intervals of the AUCROC for each cohort. Box plots represent the minimum, Q1, median, Q3, and maximum values. Gray diamonds represent the mean. P-values ​​are given by a two-sided Wilcoxon rank-sum test. p < 0.05, p < 0.01, p < 0.001, no significant difference in normal range. In this case, the relative abundance of the six UC exhaustion species was lower in patients with active UC compared to patients with inactive UC, such as... Figure 27A and 27B As shown; and compared to inactive CD, *R. oryzae*, including *Brutella ovalis*, *L. sucrase*, *Dolphobicella formate*, and *Eubacterium* species CAG: 274 were lower in patients with active CD, while some disease-enriched bacteria, such as *Germococcus mesenteriae*, were higher in active UC, and *Bacteroides fragilis*, *Escherichia coli*, and *Actinomyces* species oral taxa 181 were higher in active CD, such as... Figure 27D and 27EAs shown in the figure, this means that these bacterial species may be associated with disease activity and severity. Next, the diagnostic model was used to calculate the probability values ​​of disease risk, and no significant difference in disease probability was found between inactive and active IBD. Our model was able to distinguish inactive UC and CD patients from controls, with AUCs of 0.89 and 0.84, respectively. Figure 27C and Figure 27F As shown in the figures, these data suggest that bacterial biomarkers may not only be a result of inflammation, but may also contribute to or reflect underlying disease pathogenesis.

[0814] Validating multibacterial biomarker diagnostic models in independent cohorts

[0815] Next, we analyzed two independent datasets from Hong Kong, China (139 UCs, 139 controls; 92 CDs, 108 controls) and Australia (98 CDs, 81 controls), as well as three public datasets from the United States (53 UCs, 68 CDs, 34 controls), the Netherlands (23 UCs, 20 CDs, 22 controls), and China (25 UCs, 15 controls; 48 CDs, 54 controls).

[0816] Figures 28A-28C The figure shows the abundance and prevalence of bacterial species biomarkers and the performance of diagnostic models in cohorts from different ethnic groups and regions. Figure 28A and Figure 28B The figure shows the bacterial species biomarker characteristics for the diagnosis of UC and CD in patients and healthy individuals used in the discovery cohort, validation cohort, and three downloaded public datasets. Species abundance was normalized to log2 fold change relative to the control mean (log2FC). P-values ​​were calculated using a two-sided Wilcoxon rank-sum test. After controlling for multiple tests using Benjamini–Hochberg correction, the P-values ​​were converted to -log10 (P-value). p < 0.05, p < 0.01, p < 0.001. Universality represents the proportion of bacteria present in the UC, CD, and healthy group for each cohort. Figure 28C The figure shows the disease probabilities calculated using a random forest model among UC / CD patients and healthy controls in the discovery cohort from Hong Kong, China, the validation cohort from Hong Kong and Australia, and public datasets (US, Netherlands, and China). Box plots represent the minimum, Q1, median, Q3, and maximum values. P-values ​​were calculated using a two-sided Wilcoxon rank-sum test. Figure 28DThe figure shows the correlations among ten UC bacterial species markers. UC-depleted bacteria are marked in green, while UC-enriched bacteria are marked in yellow. Figure 28E The figure shows the correlations among nine CD bacterial species markers. CD-depleted bacteria are marked in green, while CD-enriched bacteria are marked in orange. Red grids indicate positive correlations, while blue grids indicate negative correlations. Correlation coefficients and p-values ​​are given by Spearman correlation.

[0817] The abundance and prevalence of bacterial biomarkers in independent cohorts and public datasets were consistent with those reported in the discovery cohort. Figure 28A and 28B ).

[0818] Figure 29A-29M The figure shows the performance of models with bacterial species biomarkers in distinguishing UC or CD patients from controls in independent cohorts and public datasets. Figure 29A The figure shows the performance of the model with ten selected bacterial species biomarkers for UC in a validation cohort in Hong Kong, China, in classifying UC patients versus controls. Figures 29B-29C The figure shows the performance of the model with nine CD selected bacterial species biomarkers in a validation cohort in Hong Kong and Australia in classifying CD patients versus controls. Figures 29B-29C The figure shows the performance of the model with selected bacterial species biomarkers in classifying UC or CD patients versus controls on three downloaded public datasets. Figure 29D The figure shows the performance of the model with selected bacterial species biomarkers in classifying UC patients versus controls across three downloaded public datasets.

[0819] Figure 29E The figure shows the performance of the model with selected bacterial species biomarkers in classifying CD patients versus controls on three downloaded public datasets. Figure 29F-29G The figure shows the associations between the relative abundance of biomarkers for disease group, geography, ethnicity, and bacterial species, calculated using MaAsLin2 across all IBD cohorts. Positive associations are indicated in red, while negative associations are indicated in blue. Significant associations (FDR < 0.05) are indicated by a plus sign for positive associations and a minus sign for negative associations. The false discovery rate (FDR) was calculated using Benjamin Hochberg correction. Figure 29H The figure shows the performance of the model with selected bacterial species biomarkers in classifying UC patients (n = 817) versus controls (n = 1746) across all UC validation cohorts. Figure 29I The performance of models with selected bacterial species biomarkers in classifying CD patients (n = 1065) versus controls (n = 1873) across all CD validation cohorts. Figure 29J The figure shows the model performance in distinguishing between treated and treatment-naïve UC patients and controls in two downloaded public datasets. Figure 29K The figure shows the model performance in distinguishing between treated and treatment-naïve CD patients and controls in two downloaded public datasets. Figure 29L The figure shows the model performance in distinguishing UC patients from controls on two downloaded public datasets compared to the fecal calprotectin test. Figure 29M The figure shows the model performance in distinguishing CD patients from controls on two downloaded public datasets compared to the fecal calprotectin test. The shaded area of ​​the ROC curve represents the 95% confidence interval of the AUC for each cohort.

[0820] In this example, the diagnostic model disclosed herein demonstrated the expected performance in classifying IBD from controls, with an AUC of 0.81 (95% CI: 0.76–0.86) for UC. Figure 29A As shown in the figure; and the AUC of CD in the Hong Kong, China cohort was 0.83 (95% CI: 0.77–0.89), as Figure 29B As shown in the figure; and the AUC of CD in the Australian cohort was 0.73 (95% CI: 0.65–0.80), as Figure 29C As shown in the figure. Using datasets from the United States, the Netherlands, and China, the diagnostic model achieved AUCs of 0.85, 0.87, and 0.82 for UC diagnosis, respectively. Figure 29D , Figure 28C Furthermore, for CD diagnosis, the diagnostic model achieved AUCs of 0.89, 0.86, and 0.97, respectively. Figure 29E , Figure 28C Bacterial ecological network analysis revealed co-occurrence correlations among depleted bacterial species in IBD, and stable co-exclusion correlations between disease-depleted and disease-enriched species were observed in almost all cohorts. Figures 28D-28E To further validate the diagnostic model, additional metagenomic datasets from the United States, the Netherlands, Spain, Denmark, and the United Kingdom were utilized (577 UCs, 739 CDs, and 1574 controls). By integrating data from all IBD cohorts and adjusting for geography and ethnicity, the multibacterial biomarker set remained significantly different between IBD and controls. Figure 29F-29G In all validation queues, the UC model maintained an overall AUC of 0.82, while the CD model achieved an AUC of 0.76. Figure 29H-29I In summary, these results highlight the robustness of our multibacterial biomarker set in terms of diagnostic performance across different regions and ethnicities.

[0821] Given that drugs used to induce and maintain remission in IBD can alter the diversity and composition of the gut microbiome, models were tested to determine whether they would be affected by treatment. Using metagenomic data from two public datasets (the US and the Netherlands) where treatment data for IBD cases were available, the UC model was found to distinguish treatment-naïve (n=14) and treated UC patients (n=62) from controls (n=56) with AUCs of 0.74 and 0.89, respectively, while the CD model distinguished treatment-naïve (n=20) and treated CD patients (n=66) from controls (n=56) with AUCs of 0.89 and 0.88, respectively, suggesting that the performance of the multibacterial biomarker group is unlikely to be affected by treatment. Figure 29J-29K ).

[0822] To compare the model with the commonly used IBD screening test (fecal calprotectin), data from two public datasets (US and Netherlands) were used, where fecal calprotectin data were available (46 UCs, 65 CDs, and 42 controls). The bacterial biomarker-based diagnostic model was found to have numerically higher AUC values ​​for UC (AUC 0.85 vs. 0.81) and CD (AUC 0.87 vs. 0.79) than the fecal calprotectin model. Figure 29L-29M The multibacterial biomarker group also showed higher sensitivity (72% vs. 54% for CD; 67% vs. 57% for UC) and specificity (95% vs. 86% for CD; 88% vs. 86% for UC) than fecal calprotectin. Specificity of IBD Diagnostic Models Based on Multibacterial Biomarkers

[0823] Given the common microbiome variations across various diseases, the study in this example validates the disease specificity of bacterial biomarkers. Several internal non-IBD disease metagenomic datasets from Hong Kong, China, were evaluated, including subjects with various gastrointestinal diseases (n = 439) and non-gastrointestinal diseases (n = 291). The gastrointestinal diseases included colorectal cancer (CRC, n = 160), colorectal adenoma (CA, n = 162), and irritable bowel syndrome (diarrhea subtype, IBS-D, n = 117), while the non-gastrointestinal diseases included obesity (n = 148) and cardiovascular disease (CVD; n = 143).

[0824] Figures 30A-30B The figure shows the relative abundance of bacterial species biomarkers in the UC / CD and other non-IBD disease groups in the Hong Kong, China cohort. Figure 30A The figure shows the relative abundance of 10 UC bacterial species biomarkers in UC and other non-IBD disease groups. Figure 30B The figure shows the relative abundance of biomarkers for nine CD bacterial species in CD and other non-IBD disease groups. Box plots represent the minimum, Q1, median, Q3, and maximum values. Gray diamonds represent the mean. P-values ​​were calculated using a two-tailed Wilcoxon rank-sum test. p < 0.05, p < 0.01, p < 0.001, NS was not significant. CD: Crohn's disease; UC: ulcerative colitis; IBS-D: irritable bowel syndrome (diarrhea subtype); CA: colorectal adenoma; CRC: colorectal cancer; CVD: cardiovascular disease.

[0825] Among bacterial biomarkers in UC, the depletion of Ruminococcus tortifolia and the enrichment of Clostridium spiralis are unique to UC patients compared to all other non-IBD diseases. Figure 30A Among CD bacterial biomarkers, depletion of *Roseola ovalis* and *Brutella ovalis*, and increases in *Bacteroides fragilis* and *Escherichia coli* are particularly associated with CD. Figure 30B ).

[0826] Figures 31A-31E The performance of models with bacterial species biomarkers in distinguishing patients with UC or CD from other subjects with or without gastrointestinal disorders in an international cohort was demonstrated. Figure 31A The figure shows the composition of the international multi-disease dataset from different countries and regions. Figure 31B The figure shows a comparison of disease probabilities generated by a UC model based on biomarkers of ten UC bacterial species in controls (n = 2391), CVD (n = 143), obesity (n = 318), CA (n = 230), CRC (n = 372), IBS-D (n = 146), and UC patients (n = 817). The data are presented as box plots with the median (center line), 25th and 75th percentiles (box boundaries), and 5th and 95th percentiles (whiskers). Figure 31C The figure shows the performance of the UC model in classifying UC patients (n = 817) from other non-IBD subjects (n = 3600). Figure 31DThe figure shows a comparison of disease probabilities generated by a CD model based on biomarkers of nine CD bacterial species in controls (n = 2391), CVD (n = 143), obesity (n = 318), CA (n = 230), CRC (n = 372), IBS-D (n = 146), and CD patients (n = 1065). The data are presented as box plots with the median (center line), 25th and 75th percentiles (box boundaries), and 5th and 95th percentiles (whiskers). Figure 31E The figure shows the performance of the CD model in classifying CD patients (n = 1065) from other non-IBD subjects (n = 3600). Box plots represent the minimum, Q1, median, Q3, and maximum values. P-values ​​were calculated using a two-sided Wilcoxon rank-sum test. p < 0.0001. The shaded area of ​​the ROC curve represents the 95% confidence interval of the AUC for each cohort. CD: Crohn's disease; UC: ulcerative colitis; IBS-D: irritable bowel syndrome (diarrhea subtype); CA: colorectal adenoma; CRC: colorectal cancer; CVD: cardiovascular disease.

[0827] To validate the model in a diverse international non-IBD cohort, additional metagenomic data from 843 non-IBD cohorts (212 CRC, 68 CA, 29 IBS-D, 170 obese, and 364 controls) from Austria, France, Germany, Japan, the United States, and Denmark were included. Figure 31A The disease probabilities generated by the model showed significant differences between IBD and non-IBD subjects. Figure 31B , Figure 31D The UC diagnostic model distinguishes UC patients from non-IBD subjects, with an AUC of 0.78 (). Figure 31C The CD diagnostic model distinguishes CD from non-IBD subjects with an AUC of 0.72 (). Figure 31E Overall, these results indicate that our multibacterial group is specific for UC and CD.

[0828] Development of a Universal IBD Model Based on Multibacterial Biomarkers

[0829] Since the need for a universal biomarker to distinguish between IBD and non-IBD subjects has not been met, an IBD model has been developed that uses a total of 18 bacterial species for identification of UC and CD (UC: 10 species; CD: 9 species; 1 species overlaps in UC and CD).

[0830] Figures 32A-32C The figure shows the performance of the general IBD model in classifying IBD and non-IBD subjects. Figure 32A The figure shows the ROC of the general IBD model in classifying IBD from control and non-IBD in the test set, IBD validation cohort and non-IBD cohort. Figure 32B The figure shows a typical criteria (STARD) diagram for reporting diagnostic accuracy studies, which reports the participant flow in the Independent International IBD Cohort (IBD = 1882, Control = 2027). Figure 32C The figure shows a comparison of the diagnostic performance of the universal IBD model and fecal calprotectin in classifying IBD from IBS subjects. The shaded areas of the ROC curves represent the 95% confidence intervals of the AUC for each cohort.

[0831] The IBD model achieved an AUC of 0.91 (95% CI: 0.85–0.97) in the discovery cohort, with a sensitivity of 92% and a specificity of 75%, and an AUC of 0.81 (95% CI: 0.80–0.82) in the validation cohort (controls = 2,027, IBD = 1,882), with a sensitivity of 78% and a specificity of 70%. In an international multi-disease cohort including subjects with both GI and non-GI diseases (IBD = 1,882; non-IBD = 3,600), the model distinguished between IBD and non-IBD with an AUC of 0.77 (95% CI: 0.75–0.78). Figure 32A and Figure 32B In the pilot cohort, a direct comparison was made between the mult...

Claims

1. A method for determining, diagnosing, preventing, or treating inflammatory bowel disease (IBD) in an individual, the method comprising the following steps: a) Determine the relative abundance of a set of bacterial species in a fecal sample from the individual, wherein the set of bacterial species comprises one or more bacterial species selected from the group consisting of: Actinomycetes species oral taxa 181, Bacteroides fragilis, Escherichia coli, Lawsonia sacchariformis, Eubacterium species CAG: 274, Rosette of inulin, Rosette of the gut, Oval rumenococci (Oval Brutella), Formic acid-producing Dorl, Bileophilus warwickii, Clostridium truncatella, Clostridium sacchariformis, Formic acid-producing Clostridium, Formic acid-producing Gemmaie, Visceral Osmophilus, Ruminococcus tortiformis, Clostridium spiralis, Gemmaietta esculenta, and Broutella henselae; b) The relative abundance of each bacterial species in the bacterial species set from the individual is compared with the relative abundance of each bacterial species in the bacterial species set in the reference dataset to generate a risk score; c) If the risk score is higher than the cutoff value, the individual is determined to have an increased risk of developing IBD or is currently suffering from IBD; as well as d) If it is determined that the individual has an increased risk of developing IBD or already has IBD, then the individual may optionally be treated; The IBD mentioned therein is either active or inactive.

2. A method for determining, diagnosing, preventing, or treating in an individual inflammatory bowel disease (IBD) and / or an IBD subtype, wherein the IBD subtype is ulcerative colitis (UC) or Crohn's disease (CD), the method comprising the following steps: (a) Determine the relative abundance of a first set of bacterial species, optionally the level of fecal calprotectin, and optionally the relative abundance of a second set of bacterial species in a fecal sample from the individual, wherein the first set of bacterial species and the second set of bacterial species each independently comprise one or more bacterial species selected from the group consisting of: Actinomycetes species oral taxa 181, Bacteroides fragilis, Escherichia coli, Lawsonia sucrase, Eubacterium species CAG: 274, Rosette spp., Rosette spp., Ruminococcus ovalis (Broutella ovalis), Formicotrophus dorsi, Bileophilus warwickii, Clostridium truncatella, Clostridium spp., Formicotrophus spp., Gemmaella formica, Osmotherium visceratum, Ruminococcus tortifolium, Clostridium spiralis, Gemmaella spp., and Broutella henselae; (b) The relative abundance of each bacterial species in the first bacterial species set from the individual is compared with the relative abundance of each bacterial species in the first bacterial species set in the first reference dataset, and optionally the fecal calprotectin level from the individual is compared with the fecal calprotectin level in the first reference dataset using a first machine learning model to generate a first risk score. (c) If the first risk score is higher than the first cutoff value, the individual is determined to have an increased risk of developing IBD or to have IBD. (d) If it is determined that the individual has an increased risk of developing IBD or is currently suffering from IBD, a second machine learning model may be used to compare the relative abundance of each bacterial species in the second bacterial species set from the individual with the relative abundance of each bacterial species in the second bacterial species set in the second reference dataset to generate a second risk score. (e) Optionally, if the second risk score is higher than the second cutoff value, the individual is determined to have an increased risk of developing CD or to have CD, or if the risk score is lower than or equal to the second cutoff value, the individual is determined to have an increased risk of developing UC or to have UC. as well as (f) If it is determined in step (c) that the individual has an increased risk of developing IBD or is currently suffering from IBD, or if steps (d) and (e) are performed and it is determined that the individual has an increased risk of developing UC or CD or is currently suffering from UC or CD, then the individual may optionally be treated.

3. The method of claim 2, wherein the first set of bacterial species comprises one or more bacterial species selected from the group consisting of: Actinomycetes species oral taxa 181, Escherichia coli, Lawsonia sucrase, Eubacterium species CAG: 274, Enterobacter rosenbergii, Ruminococcus ovalis (Broutella ovalis), Bileophilus warwickii, Clostridium truncatella, Clostridium sacchariformis, Gemmaformis, Osmotherium visceratum, Ruminococcus tortiformis, Clostridium spiralis, Gemmaella esculenta, and Brooutella henselae.

4. The method of claim 2 or claim 3, wherein the first set of bacterial species comprises two or more bacterial species, wherein the two or more bacterial species comprise two bacterial species selected from the group consisting of: *Bacillus viscerans* and *Sclerotium mesenteriae*; *Bacillus viscerans* and *Clostridium velutipes*; *Bacillus viscerans* and *Brutella hensenoides*; *Bacillus viscerans* and *Escherichia coli*; *Bacillus viscerans* and *Actinomyces* species oral taxonomy group 181; *Sclerotium mesenteriae* and *Clostridium velutipes*; *Sclerotium mesenteriae* and *Ruminococcus ovalis* (*Brutella ovalis*); *Sclerotium mesenteriae* and *Brutella hensenoides*; *Sclerotium mesenteriae* and *Clostridium velutipes*; and *Sclerotium mesenteriae* and *Escherichia coli*.

5. The method of claim 2 or claim 3, wherein the first set of bacterial species comprises three or more bacterial species, wherein the three or more bacterial species comprise a group of three bacterial species selected from the group consisting of: *Ostomyces visceratus*, *Twincocephalus mesenteriae*, and *Clostridium velutipes*; *Ostomyces visceratus*, *Twincocephalus mesenteriae*, and *Brutella hensenoides*; *Ostomyces visceratus*, *Twincocephalus mesenteriae*, and *Clostridium spiralis*; *Twincocephalus mesenteriae*, *Clostridium velutipes*, and *Ruminococcus ovalis* (*Brutella ovalis*); *Twincocephalus mesenteriae* Cocci, *Brutella ovalis* and *Brutella hensenii*; *Twincocephalus measlese*, *Brutella ovalis* and *Clostridium spiralis*; *Twincocephalus measlese*, *Brutella ovalis* and *Escherichia coli*; *Clostridium tenuis*, *Brutella ovalis* and *Brutella hensenii*; *Clostridium tenuis*, *Brutella ovalis* and *Clostridium spiralis*; and *Clostridium tenuis*, *Brutella ovalis* and *Escherichia coli*.

6. The method of claim 2 or claim 3, wherein the first set of bacterial species comprises two or more bacterial species selected from the group consisting of: visceral bacterium, measles twin cocci, Clostridium tumefaciens, ovale broutococci (ovale broutococci), and broutococci hensenii.

7. The method of claim 2 or claim 3, wherein the first set of bacterial species comprises at least one bacterial species selected from the group consisting of: visceral bacterium, measles twin cocci, Clostridium tumefaciens and ovoid rumenococci (ovoid broutella).

8. The method of claim 2 or claim 3, wherein the first set of bacterial species comprises at least one bacterial species selected from the group consisting of: visceral bacterium, measles twin cocci, Clostridium tumefaciens, ovale rumenococci (ovale Brutella) and bilirubinophilus.

9. The method of claim 2 or claim 3, wherein the first set of bacterial species comprises at least one bacterial species selected from the group consisting of: visceral bacterium, measles twin cocci, Clostridium truncatella, Brontë hanseniella, ovale rumenococci (ovale broutuli), and Clostridium spiralis.

10. The method of claim 2 or claim 3, wherein the first set of bacterial species comprises at least three bacterial species selected from the group consisting of: visceral bacterium, measles twin cocci, Clostridium truncatella, Brontë hanseniella, ovale rumenococci (ovale broutuli), and Clostridium spiralis.

11. The method of claim 2 or claim 3, wherein the first set of bacterial species comprises at least three bacterial species selected from the group consisting of: visceral bacterium, measles twin cocci, Clostridium truncatum, Broutella hensenae, Escherichia coli, and Clostridium spiralis.

12. The method of any one of claims 2-11, wherein the second set of bacterial species comprises two or more bacterial species, wherein the two or more bacterial species comprise two bacterial species selected from the group consisting of: *Brutella ovalis* and *G. formic acid*; *Brutella ovalis* and *R. rostrella*; *Brutella ovalis* and *Brutella hensenii*; *Brutella ovalis* and *R. rostrella inulin*; *Brutella ovalis* and *Germococcus mesenteriae*; *G. formic acid* and *Brutella hensenii*; *G. formic acid* and *R. rostrella*; and *G. formic acid* and *Germococcus mesenteriae*.

13. The method of any one of claims 2-11, wherein the second set of bacterial species comprises at least one bacterial species selected from the group consisting of: ovomycosis (Brutella ovomycosis), formic acid gimida, enterorosiformis, Broutella hensenae, Gemmaella mesenteriae, and Bacteroides fragilis.

14. The method of any one of claims 2-11, wherein the second set of bacterial species comprises at least one bacterial species selected from the group consisting of: ovomycosis (Brutella ovalis), formic acid-producing Gemmaieux, Enterobacter rosenbergii, Brautella hensenae, Gemmaieux var. measlese, Lawsonia fragilis, Bacteroides fragilis, and Escherichia coli.

15. The method of any one of claims 2-11, wherein the second set of bacterial species comprises at least one bacterial species selected from the group consisting of: ovomycosis (Brutella ovomycosis), formic acid-producing Gemmaieux, enteric Rosebes, Brautella hensenoides, Gemmaieux var. measles, Lawsonia solani, Bacteroides fragilis, Escherichia coli, Clostridium truncatum, and Osmidrosis viscera.

16. The method of any one of claims 2-15, wherein the first set of bacterial species comprises four or more bacterial species selected from the group consisting of: Actinomycetes oral taxa 181, Escherichia coli, Lawsonia solani, Eubacterium species CAG: 274, Enterobacter rosenbergii, Ruminococcus ovalis (Broutella ovalis), Biliophilus warwickii, Clostridium truncatella, Clostridium sacchariformis, Formic acid-producing Gemmaieux, Ostomyces viscerans, Ruminococcus tortiformis, Clostridium spiralis, Gemella mesenteriae, and Broutella hensenii, and the second set of bacterial species comprises four or more bacterial species selected from the group consisting of: Actinomycetes oral taxa 181, Bacteroides fragilis, Escherichia coli, Lawsonia solani, Eubacterium species CAG: 274, 274. Inulin-eating Roselle, Enteric Roselle, Oval-shaped Ruminococcus (Oval Broutella), Formic acid-producing Dorl, Bileophilus warwickii, Clostridium truncatella, Clostridium sacchariformis, Formic acid-producing Gemma, Visceral Odorbacterium, Twisted Ruminococcus, Clostridium spiralis, Measles twin cocci, and Broutella hensenoides.

17. The method as claimed in any of the preceding claims, wherein the first set of bacterial species comprises substantially the following: The oral taxonomy of Actinomycetes species 181, *Escherichia coli*, *Lawsonia unagalactiae*, *Eubacterium* species CAG: 274, *Roseidonella enterica*, *Broutella ovalis*, *Biliobacterium warfarinii*, *Clostridium truncatella*, *Clostridium sacchariformis*, *Gemmelella formica*, *Ostomyces viscerans*, *Ruminococcus tortifolia*, *Clostridium spiralis*, *Twincocephala spp.*, and *Broutella henselae*, and the second bacterial species set is basically composed of the following: oral taxonomy of Actinomycetes species 181, *Bacteroides fragilis*, *Escherichia coli*, *Lawsonia unagalactiae*, *Eubacterium* species CAG:

274. Inulin-eating Roselle, Enteric Roselle, Oval-shaped Ruminococcus (Oval Broutella), Formic acid-producing Dorl, Bileophilus warwickii, Clostridium truncatella, Clostridium sacchariformis, Formic acid-producing Gemma, Visceral Odorbacterium, Twisted Ruminococcus, Clostridium spiralis, Measles twin cocci, and Broutella hensenoides.

18. The method as described in any of the preceding claims, wherein The first reference dataset was obtained by determining the relative abundance of each bacterial species in the first bacterial species set and, optionally, the level of fecal calprotectin in fecal samples from a first subject group, which included patients diagnosed with IBD and individuals without IBD in a reference cohort; and The second reference dataset was obtained by determining the relative abundance of each bacterial species in the second bacterial species set in fecal samples from a second subject group, which included patients diagnosed with UC and patients diagnosed with CD.

19. The method of claim 18, wherein the subjects in the second subject group are selected from patients in the first subject group who have been diagnosed with IBD.

20. The method as described in any of the preceding claims, wherein the first machine learning model is constructed by: i) The first reference dataset is obtained by determining the relative abundance of each bacterial species in the first bacterial species set and the optional fecal calprotectin level in fecal samples from the first subject group, which includes patients diagnosed with IBD and individuals in the reference cohort who do not have IBD. ii) Divide the first reference dataset into a first training dataset and a first test dataset; iii) Divide the first training dataset into a first training subset and a first validation subset; iv) Train the first candidate model using the first training subset and evaluate the performance of the first candidate model using the first validation subset; v) Optionally, the first candidate model can be tuned by adjusting the hyperparameters in the first candidate model; vi) Repeat steps (iv) and optional steps (v) to generate a first candidate model set and select the best first candidate model from the first candidate model set; vii) Evaluate the performance of the best first candidate model using the first test dataset; viii) Optionally, using the hyperparameter set corresponding to the best first candidate model, a first final model is trained with the first reference dataset to produce the first final model; as well as ix) Use the best first candidate model from step (vii), or if the first final model was trained using step (viii), use the first final model from step (viii) as the first machine learning model for the comparison step in step (b) to generate the first risk score.

21. The method as described in any of the preceding claims, wherein the second machine learning model is constructed by: i. The second reference dataset is obtained by determining the relative abundance of each bacterial species in the second bacterial species set from fecal samples from the second subject group, which includes patients diagnosed with UC and patients diagnosed with CD; ii. Divide the second reference dataset into a second training dataset and a second test dataset; iii. Divide the second training dataset into a second training subset and a second validation subset; iv. Train the second candidate model using the second training subset and evaluate the performance of the second candidate model using the second validation subset; v. Optionally, the second candidate model can be tuned by adjusting the hyperparameters in the second candidate model; vi. Repeat steps (iv) and optional steps (v) to generate a second set of candidate models, and select the best second candidate model from the second set of candidate models; vii. Evaluate the performance of the best second candidate model using the second test dataset; viii. Using the hyperparameter set corresponding to the best second candidate model, train the second final model with the second reference dataset to produce the second final model; and ix. Use the best second candidate model from step (vii), or if the second final model was trained using step (viii), use the second final model from step (viii) as the second machine learning model for the comparison step in step (d) to generate the second risk score.

22. The method as claimed in any of the preceding claims, wherein the first machine learning model is a first random forest model, and wherein the first risk score in step (b) is generated by the following steps: (b1) Using the relative abundance of the first bacterial species set in the first reference dataset and the optional fecal calprotectin level, generate a set of decision trees using the first random forest model; and (b2) Run the relative abundance of the first bacterial species set from the individuals and the optional fecal calprotectin level along the decision tree set to generate the first risk score.

23. The method as claimed in any of the preceding claims, wherein the second machine learning model is a second random forest model, and wherein the second risk score in step (d) is generated by the following steps: (d1) Using the relative abundance of the second bacterial species set in the second reference dataset, generate a set of decision trees using the second random forest model; and (d2) Run the relative abundance of the second set of bacterial species from the individual along the decision tree set to generate the second risk score.

24. The method of claim 20, wherein the best first candidate model is the first candidate model with the highest performance in the first candidate model set.

25. The method of claim 21, wherein the best second candidate model is the second candidate model with the highest performance in the second candidate model set.

26. The method of any one of claims 20-25, wherein the performance of the first candidate model and / or the second candidate model is evaluated by the area under the curve (AUC) of the receiver operating characteristic (ROC) curve of the first candidate model and / or the second candidate model.

27. The method as claimed in any of the preceding claims, wherein the first cutoff value and / or the second cutoff value are in the range of 0.4 to 0.

75.

28. The method as described in any of the preceding claims, wherein the first cutoff value is determined by the following steps: (c1) Calculate the Yoden Index using the first risk score data obtained from the first reference dataset; and (c2) Determine the first cutoff value based on the Yoden index.

29. The method as described in any of the preceding claims, wherein the second cutoff value is determined by the following steps: (e1) Calculate the Yoden Index using the second risk score data obtained from the second reference dataset; and (e2) Determine the second cutoff value based on the Yoden index.

30. The method as claimed in any of the preceding claims, wherein the IBD is an active IBD or an inactive IBD, the UC is an active UC or an inactive UC, and the CD is an active CD or an inactive CD.

31. The method of any of the preceding claims, wherein the method identifies the individual’s risk of having IBD relative to non-IBD diseases, wherein the non-IBD diseases include colorectal cancer, colorectal adenoma, irritable bowel syndrome (IBS), obesity, cardiovascular disease, acid reflux, colonic polyps, functional dyspepsia, and gastrointestinal stromal tumors (GIST).

32. The method of any of the preceding claims, wherein the method identifies an individual as having an increased risk of developing inactive IBD relative to irritable bowel syndrome (IBS) or is currently suffering from inactive IBD.

33. The method of any of the preceding claims, wherein the relative abundance of the first bacterial species set and the relative abundance of the second bacterial species set are determined by metagenomic sequencing and / or polymerase chain reaction (PCR).

34. The method of claim 33, wherein the PCR is a digital droplet PCR (ddPCR) or quantitative PCR (qPCR) performed using one or more primer sets and optionally one or more fluorophore-quencher probes, wherein each primer set is used to amplify a target nucleic acid sequence of a bacterial species belonging to the first bacterial species set or the second bacterial species set to produce an amplification product, and each fluorophore-quencher probe is used to detect the amplification product of each primer set.

35. The method of claim 34, wherein the target nucleic acid sequence amplified by each primer set comprises a portion of a nucleotide sequence selected from the group consisting of: SEQ ID NO: 76, SEQ ID NO: 77, SEQ ID NO: 78, SEQ ID NO: 80, SEQ ID NO: 81, SEQ ID NO: 82, SEQ ID NO: 85, SEQ ID NO: 89, SEQ ID NO: 92, SEQ ID NO: 137, SEQ ID NO: 139, SEQ ID NO: 141, SEQ ID NO: 143, SEQ ID NO: 145, SEQ ID NO: 147, SEQ ID NO: 160, SEQ ID NO: 162, and SEQ ID NO:

164.

36. The method of claim 34, wherein the target nucleic acid sequence amplified by each primer set comprises a nucleotide sequence selected from the group consisting of: SEQ ID NO: 58, SEQ ID NO: 59, SEQ ID NO: 60, SEQ ID NO: 62, SEQ ID NO: 63, SEQ ID NO: 64, SEQ ID NO: 67, SEQ ID NO: 71, SEQ ID NO: 74, SEQ ID NO: 136, SEQ ID NO: 138, SEQ ID NO: 140, SEQ ID NO: 142, SEQ ID NO: 144, SEQ ID NO: 146, SEQ ID NO: 159, SEQ ID NO: 161 and SEQ ID NO:

163.

37. The method of claim 34, wherein the one or more primer sets comprise the forward primer of SEQ ID NO: 1 and the reverse primer of SEQ ID NO: 2; the forward primer of SEQ ID NO: 3 and the reverse primer of SEQ ID NO: 4; the forward primer of SEQ ID NO: 5 and the reverse primer of SEQ ID NO: 6; the forward primer of SEQ ID NO: 9 and the reverse primer of SEQ ID NO: 10; the forward primer of SEQ ID NO: 11 and the reverse primer of SEQ ID NO: 12; the forward primer of SEQ ID NO: 13 and the reverse primer of SEQ ID NO: 14; the forward primer of SEQ ID NO: 19 and the reverse primer of SEQ ID NO: 20; the forward primer of SEQ ID NO: 33 and the reverse primer of SEQ ID NO: 34; the forward primer of SEQ ID NO: 94 and the reverse primer of SEQ ID NO: 95; the forward primer of SEQ ID NO: 96 ...5; the reverse primer of SEQ ID NO: 96; the forward primer of SEQ ID NO: 94 and the reverse primer of SEQ ID NO: 95; the forward primer of SEQ ID NO: 96 and the reverse primer of SEQ ID NO: 94; the reverse primer of SEQ ID NO: 95; the forward primer of SEQ ID NO: 96; the reverse primer of SEQ ID NO: 94; the reverse primer of SEQ ID NO: 95; the forward primer of SEQ ID NO: 96; the reverse primer of SEQ ID NO: 94; the reverse primer of SEQ ID NO: 95; the forward primer of SEQ ID NO: 96; the reverse primer of The reverse primer of SEQ ID NO: 97; the forward primer of SEQ ID NO: 98 and the reverse primer of SEQ ID NO: 99; the forward primer of SEQ ID NO: 100 and the reverse primer of SEQ ID NO: 101; the forward primer of SEQ ID NO: 102 and the reverse primer of SEQ ID NO: 103; the forward primer of SEQ ID NO: 104 and the reverse primer of SEQ ID NO: 105; the forward primer of SEQ ID NO: 106 and the reverse primer of SEQ ID NO: 107; the forward primer of SEQ ID NO: 118 and the reverse primer of SEQ ID NO: 119; and / or the forward primer of SEQ ID NO: 120 and the reverse primer of SEQ ID NO: 121; and each fluorophore-quencher probe comprises a probe sequence selected from the group consisting of: SEQ ID NO: 39, SEQ ID NO: 40, SEQ ID NO: 41, SEQ ID NO: 97, SEQ ID NO: 98, SEQ ID NO: 99, SEQ ID NO: 99, SEQ ID NO: 100, SEQ ID NO: 101, SEQ ID NO: 102, SEQ ID NO: 103, SEQ ID NO: 104, SEQ ID NO: 105, SEQ ID NO: 106, SEQ ID NO: 107, SEQ ID NO: 118, SEQ ID NO: 119, and / or SEQ ID NO: 120, SEQ ID NO: 121; and each fluorophore-quencher probe comprises a probe sequence selected from the group consisting of: SEQ ID NO: 39, SEQ ID NO: 40, SEQ ID NO: 41, SEQ ID NO: 99, SEQ ID NO: 99, SEQ ID NO: 99, SEQ ID NO: 99, SEQ ID NO: 99 NO: 43, SEQ ID NO: 44, SEQ ID NO: 45, SEQ ID NO: 48, SEQ ID NO: 52, SEQ ID NO: 55, SEQ ID NO: 122, SEQ ID NO: 123, SEQ ID NO: 124, SEQ ID NO: 125, SEQ ID NO: 126, SEQ ID NO: 127, SEQ ID NO: 128, SEQ ID NO: 134 and SEQ ID NO:

135.

38. The method of any of the preceding claims, wherein if the individual is determined to have an increased risk of developing IBD, UC, or CD, or is currently suffering from IBD, UC, or CD, the individual is treated with therapeutic therapy, fecal microbiota transplantation (FMT), intestinal microbiota transplantation (IMT), lifestyle and dietary changes, or surgical intervention.

39. The method of claim 38, wherein the therapeutic treatment is selected from the group consisting of 5-aminosalicylic acid (5-ASA), steroids, thiopurines, TNF inhibitors, anti-integrins, and anti-interleukin (IL) 12 / 23.

40. The method of claim 39, wherein the 5-aminosalicylate is mesalazine or sulfasalazine; wherein the steroid is budesonide or prednisolone; wherein the thiopurine is azathioprine, mercaptopurine, or methotrexate; wherein the TNF inhibitor is infliximab, adalimumab, or cetuzumab; wherein the anti-integrin is vedolizumab; and wherein the anti-interleukin (IL) 12 / 23 is ustekinumab.

41. The method of any of the preceding claims, wherein if the individual is determined to have an increased risk of developing IBD, UC, or CD, or is currently suffering from IBD, UC, or CD, the individual is treated with appropriate IBD treatment, UC treatment, or CD treatment, respectively.

42. The method of claim 41, wherein the UC treatment is selected from the group consisting of 5-aminosalicylate (5-ASA), tofacitinib, mijizumab, and gusecubitumab.

43. The method of claim 41, wherein the CD treatment is nastatinumab.

44. The method of claim 41, wherein the UC treatment is a composition comprising one or more bacterial species selected from the group consisting of: *Clostridium sacchariformis*, *Clostridium tenuifolium*, *Gemmelella formica*, *Ruminococcus tortifolia*, *Ostomyces viscerans*, *Collagenia fecalis*, *Bacillus sacchariformis*, *Collinus faecium*, *Vibrio vulnificus* species CAG: 241, *Lawsonia sacchariformis*, *Toxococcus butyricum*, *Clostridium* species CAG: 58, *Eubacterium* species CAG: 274, and *Bacillus warfarinia*, and optionally a pharmaceutically acceptable carrier.

45. The method of claim 41, wherein the CD treatment is a composition comprising one or more bacterial species selected from the group consisting of: *Ruminococcus ovalis* (Brutella ovalis), *Lawsonia ovalis* (unsaturated), *Roseobacterium inulinum*, *Dolphobicella formate*, *Eubacterium tumefaciens*, *Roseobacterium tumefaciens*, *Eubacterium tumefaciens*, *Anaerobic Bacillus hardruciformis*, *Vibrio vulnificus* species CAG: 241, *Vibrio vulnificus* species 57_20, *Lactobacillus rhamnosus*, *Eubacterium* species CAG: 274, and *Romebutzella ileum*, and optionally a pharmaceutically acceptable carrier.

46. ​​The method as claimed in any of the preceding claims, wherein the first machine learning model and / or the second machine learning model is a generalized linear model.

47. The method of any of the preceding claims, wherein the individual is an individual who has not received IBD treatment, an IBD treatment-naïve individual, or an individual exposed to IBD treatment.

48. A method for diagnosing, preventing, or treating an IBD subtype in an individual, wherein the IBD subtype is ulcerative colitis (UC) or Crohn's disease (CD), the method comprising the steps of: a) Determine whether the individual has an increased risk of developing IBD or currently has IBD; b) If it is determined that the individual has an increased risk of developing IBD or is currently suffering from IBD, the relative abundance of a second set of bacterial species in a fecal sample from the individual is determined, wherein the second set of bacterial species comprises one or more bacterial species selected from the group consisting of: Actinomycete species oral taxa 181, Bacteroides fragilis, Escherichia coli, Lawsonia sacchariformis, Eubacterium species CAG: 274, Rosette of inulin, Rosette of the gut, Oval rumenococcus (Oval Brutella), Formic acid-producing Dorl, Bileophilus warwickii, Clostridium truncatella, Clostridium sacchariformis, Formic acid-producing Gastroenteritis, Visceral Odorobacterium, Ruminococcus tortiformis, Clostridium spiralis, Gemmaella muscarinae, and Broutella hensenii; c) Using a second machine learning model, the relative abundance of each bacterial species in the second bacterial species set from the individual is compared with the relative abundance of each bacterial species in the second bacterial species set in the second reference dataset to generate a second risk score; d) If the second risk score is higher than the second cutoff value, the individual is determined to have an increased risk of developing CD or to currently have CD; or if the risk score is lower than or equal to the second cutoff value, the individual is determined to have an increased risk of developing UC or to currently have UC; and e) If it is determined that the individual has an increased risk of developing UC or CD, or is already suffering from UC or CD, the individual may optionally be treated.

49. The method of claim 48, wherein the determining step in step (a) further comprises the following steps: (a1) Determine the relative abundance of a first set of bacterial species and, optionally, the level of fecal calprotectin in a fecal sample from said individual, wherein said first set of bacterial species comprises one or more bacterial species selected from the group consisting of: Actinomycetes species oral taxa 181, Bacteroides fragilis, Escherichia coli, Lawsonia sucrase, Eubacterium species CAG:274, Rosette of inulin, Rosette of the gut, Oval rumenococci (Oval Brutella), Formic acid-producing Dorl, Bileophilus warwickii, Clostridium truncatella, Clostridium sucrase, Formic acid-producing G. formic acid, Osmophilus viscera, Ruminococcus tortiformis, Clostridium spiralis, Gemmaella muscarinae, and Broutella hensenae; (a2) The relative abundance of each bacterial species in the first bacterial species set from the individual is compared with the relative abundance of each bacterial species in the first bacterial species set in the first reference dataset, and optionally the fecal calprotectin level from the individual is compared with the fecal calprotectin level in the first reference dataset using a first machine learning model to generate a first risk score. (a3) If the first risk score is higher than the first cutoff value, the individual is determined to have an increased risk of IBD or is currently suffering from IBD.

50. The method of claim 48 or claim 49, wherein the second set of bacterial species comprises two or more bacterial species, wherein the two or more bacterial species comprise two bacterial species selected from the group consisting of: *Brutella ovalis* and *G. formica*; *Brutella ovalis* and *R. rostrella*; *Brutella ovalis* and *Brutella hensenii*; *Brutella ovalis* and *R. rostrella*; *Brutella ovalis* and *G. rostrella*; *G. formica* and *Brutella hensenii*; *G. formica* and *R. rostrella*; and *G. formica* and *G. rostrella*.

51. The method of claim 48 or claim 49, wherein the second set of bacterial species comprises at least one bacterial species selected from the group consisting of: ovomycosis (Brutella ovalis), formic acid-producing Gemmaieux, Enterobacter rosenbergii, Broutella hensenae, Gemmaieux rubrum, and Bacteroides fragilis.

52. The method of claim 48 or claim 49, wherein the second set of bacterial species comprises at least one bacterial species selected from the group consisting of: ovomycosis (Brutella ovalis), formic acid-producing Gemmaieux, Enterobacter rosenbergii, Brautella hensenae, Gemmaieux malariae, Lawsonia sacchariformis, Bacteroides fragilis, and Escherichia coli.

53. The method of claim 48 or claim 49, wherein the second set of bacterial species comprises at least one bacterial species selected from the group consisting of: ovomycosis (Broutella ovalis), formic acid-producing Gemmaieux, enterorosiformis, Brontotherium hensenoides, Gemella muscarinica, Lawsonia solani, Bacteroides fragilis, Escherichia coli, Clostridium truncatum, and Osmidrosis viscera.

54. The method of any of the preceding claims, wherein the second set of bacterial species comprises four or more bacterial species selected from the group consisting of: Actinomycetes species oral taxa 181, Bacteroides fragilis, Escherichia coli, Lawsonia sucrase, Eubacterium species CAG: 274, Rosette of inulin, Rosette of the gut, Oval rumenococcus (Oval Brutella), Formic acid-producing Dorl, Bileophilus warwickii, Clostridium truncatella, Clostridium sucrase, Formic acid-producing G. formic acid, Gemmaella formica, Visceral bacterium, Ruminococcus tortiformis, Clostridium spiralis, Gemmaella muscarinae, and Broutella hensenae.

55. The method as described in any of the preceding claims, wherein the second set of bacterial species comprises substantially the following: Actinomycetes species oral taxonomy group 181, Bacteroides fragilis, Escherichia coli, Lawsonia sacchariformis, Eubacterium species CAG: 274, Rosette sacchariformis, Rosette sacchariformis, Brontë ovale, Formico-producing Dorl, Bileophilus warwickii, Clostridium truncatella, Clostridium sacchariformis, Formico-producing Clostridium, Gemite bacilli, Visceral bacterium, Ruminococcus tortiformis, Clostridium spiralis, Gemite bacilli, and Brontë hansenii.

56. The method as described in any of the preceding claims, wherein The second reference dataset was obtained by determining the relative abundance of each bacterial species in the second bacterial species set in fecal samples from a second subject group, which included patients diagnosed with UC and patients diagnosed with CD.

57. The method of claim 49, wherein the first reference dataset is obtained by determining the relative abundance of each bacterial species in the first bacterial species set and optionally the level of fecal calprotectin in fecal samples from a first subject group, the first subject group comprising patients diagnosed with IBD and individuals without IBD in a reference cohort; and wherein the subjects in the second subject group are selected from the patients diagnosed with IBD in the first subject group.

58. The method as described in any of the preceding claims, wherein the second machine learning model is constructed by: i. The second reference dataset is obtained by determining the relative abundance of each bacterial species in the second bacterial species set from fecal samples from the second subject group, which includes patients diagnosed with UC and patients diagnosed with CD; ii. Divide the second reference dataset into a second training dataset and a second test dataset; iii. Divide the second training dataset into a second training subset and a second validation subset; iv. Train the second candidate model using the second training subset and evaluate the performance of the second candidate model using the second validation subset; v. Optionally, the second candidate model can be tuned by adjusting the hyperparameters in the second candidate model; vi. Repeat steps (iv) and optional steps (v) to generate a second set of candidate models, and select the best second candidate model from the second set of candidate models; vii. Evaluate the performance of the best second candidate model using the second test dataset; viii. Using the hyperparameter set corresponding to the best second candidate model, train the second final model with the second reference dataset to produce the second final model; and ix. Use the best second candidate model from step (vii), or if the second final model was trained using step (viii), use the second final model from step (viii) as the second machine learning model for the comparison step in step (c) to generate the second risk score.

59. The method as claimed in any of the preceding claims, wherein the second machine learning model is a second random forest model, and wherein the second risk score in step (c) is generated by the following steps: (c1) Using the relative abundance of the second bacterial species set in the second reference dataset, generate a set of decision trees using the second random forest model; and (c2) Run the relative abundance of the second set of bacterial species from the individuals along the decision tree set to generate the second risk score.

60. The method of claim 58, wherein the best second candidate model is the second candidate model with the highest performance in the second candidate model set.

61. The method of any one of claims 58-60, wherein the performance of the second candidate model is evaluated by the area under the curve (AUC) of the receiver operating characteristic (ROC) curve of the second candidate model.

62. The method as described in any of the preceding claims, wherein the second cutoff value is in the range of 0.4 to 0.

75.

63. The method of any of the preceding claims, wherein the second cutoff value is determined by the following steps: (e1) Calculate the Yoden Index using the second risk score data obtained from the second reference dataset; and (e2) Determine the second cutoff value based on the Yoden index.

64. The method as claimed in any of the preceding claims, wherein the IBD is an active IBD or an inactive IBD, the UC is an active UC or an inactive UC, and the CD is an active CD or an inactive CD.

65. The method of any of the preceding claims, wherein the relative abundance of the first bacterial species set and / or the second bacterial species set is determined by metagenomic sequencing and / or polymerase chain reaction (PCR).

66. The method of claim 65, wherein the PCR is a digital droplet PCR (ddPCR) or quantitative PCR (qPCR) performed using one or more primer sets and optionally one or more fluorophore-quencher probes, wherein each primer set is used to amplify a target nucleic acid sequence of a bacterial species belonging to the first bacterial species set or the second bacterial species set to produce an amplification product, and each fluorophore-quencher probe is used to detect the amplification product of each primer set.

67. The method of claim 66, wherein the target nucleic acid sequence amplified by each primer set comprises a portion of a nucleotide sequence selected from the group consisting of: SEQ ID NO: 76, SEQ ID NO: 77, SEQ ID NO: 78, SEQ ID NO: 80, SEQ ID NO: 81, SEQ ID NO: 82, SEQ ID NO: 85, SEQ ID NO: 89, SEQ ID NO: 92, SEQ ID NO: 137, SEQ ID NO: 139, SEQ ID NO: 141, SEQ ID NO: 143, SEQ ID NO: 145, SEQ ID NO: 147, SEQ ID NO: 160, SEQ ID NO: 162, and SEQ ID NO:

164.

68. The method of claim 66, wherein the target nucleic acid sequence amplified by each primer set comprises a nucleotide sequence selected from the group consisting of: SEQ ID NO: 58, SEQ ID NO: 59, SEQ ID NO: 60, SEQ ID NO: 62, SEQ ID NO: 63, SEQ ID NO: 64, SEQ ID NO: 67, SEQ ID NO: 71, SEQ ID NO: 74, SEQ ID NO: 136, SEQ ID NO: 138, SEQ ID NO: 140, SEQ ID NO: 142, SEQ ID NO: 144, SEQ ID NO: 146, SEQ ID NO: 159, SEQ ID NO: 161 and SEQ ID NO:

163.

69. The method of claim 66, wherein the one or more primer sets comprise the forward primer of SEQ ID NO: 1 and the reverse primer of SEQ ID NO: 2; the forward primer of SEQ ID NO: 3 and the reverse primer of SEQ ID NO: 4; the forward primer of SEQ ID NO: 5 and the reverse primer of SEQ ID NO: 6; the forward primer of SEQ ID NO: 9 and the reverse primer of SEQ ID NO: 10; the forward primer of SEQ ID NO: 11 and the reverse primer of SEQ ID NO: 12; the forward primer of SEQ ID NO: 13 and the reverse primer of SEQ ID NO: 14; the forward primer of SEQ ID NO: 19 and the reverse primer of SEQ ID NO: 20; the forward primer of SEQ ID NO: 33 and the reverse primer of SEQ ID NO: 34; the forward primer of SEQ ID NO: 94 and the reverse primer of SEQ ID NO: 95; the forward primer of SEQ ID NO: 96 and the reverse primer of SEQ ID NO: 95; the forward primer of SEQ ID NO: 96 and the reverse primer of SEQ ID NO: 94; the reverse primer of SEQ ID NO: 95; the forward primer of SEQ ID NO: 96; ... The reverse primer of SEQ ID NO: 97; the forward primer of SEQ ID NO: 98 and the reverse primer of SEQ ID NO: 99; the forward primer of SEQ ID NO: 100 and the reverse primer of SEQ ID NO: 101; the forward primer of SEQ ID NO: 102 and the reverse primer of SEQ ID NO: 103; the forward primer of SEQ ID NO: 104 and the reverse primer of SEQ ID NO: 105; the forward primer of SEQ ID NO: 106 and the reverse primer of SEQ ID NO: 107; the forward primer of SEQ ID NO: 118 and the reverse primer of SEQ ID NO: 119; and / or the forward primer of SEQ ID NO: 120 and the reverse primer of SEQ ID NO: 121; and each fluorophore-quencher probe comprises a probe sequence selected from the group consisting of: SEQ ID NO: 39, SEQ ID NO: 40, SEQ ID NO: 41, SEQ ID NO: 97, SEQ ID NO: 98, SEQ ID NO: 99, SEQ ID NO: 99, SEQ ID NO: 100, SEQ ID NO: 101, SEQ ID NO: 102, SEQ ID NO: 103, SEQ ID NO: 104, SEQ ID NO: 105, SEQ ID NO: 106, SEQ ID NO: 107, SEQ ID NO: 118, SEQ ID NO: 119, and / or SEQ ID NO: 120, SEQ ID NO: 121; and each fluorophore-quencher probe comprises a probe sequence selected from the group consisting of: SEQ ID NO: 39, SEQ ID NO: 40, SEQ ID NO: 41, SEQ ID NO: 99, SEQ ID NO: 99, SEQ ID NO: 99, SEQ ID NO: 99, SEQ ID NO: 99 NO: 43, SEQ ID NO: 44, SEQ ID NO: 45, SEQ ID NO: 48, SEQ ID NO: 52, SEQ ID NO: 55, SEQ ID NO: 122, SEQ ID NO: 123, SEQ ID NO: 124, SEQ ID NO: 125, SEQ ID NO: 126, SEQ ID NO: 127, SEQ ID NO: 128, SEQ ID NO: 134 and SEQ ID NO:

135.

70. The method of any of the preceding claims, wherein if the individual is determined to have an increased risk of developing IBD, UC, or CD, or is currently suffering from IBD, UC, or CD, the individual is treated with therapeutic therapy, fecal microbiota transplantation (FMT), intestinal microbiota transplantation (IMT), lifestyle and dietary changes, or surgical intervention.

71. The method of claim 70, wherein the therapeutic treatment is selected from the group consisting of 5-aminosalicylate (5-ASA), steroids, thiopurines, TNF inhibitors, anti-integrins, and anti-interleukin (IL) 12 / 23.

72. The method of claim 71, wherein the 5-aminosalicylate is mesalazine or sulfasalazine; wherein the steroid is budesonide or prednisolone; wherein the thiopurine is azathioprine, mercaptopurine, or methotrexate; wherein the TNF inhibitor is infliximab, adalimumab, or cetuzumab; wherein the anti-integrin is vedolizumab; and wherein the anti-interleukin (IL) 12 / 23 is ustekinumab.

73. The method of any of the preceding claims, wherein if the individual is determined to have an increased risk of developing IBD, UC, or CD, or is currently suffering from IBD, UC, or CD, the individual is treated with appropriate IBD treatment, UC treatment, or CD treatment, respectively.

74. The method of claim 73, wherein the UC treatment is selected from the group consisting of 5-aminosalicylate (5-ASA), tofacitinib, mijizumab, and gusecubitumab.

75. The method of claim 73, wherein the CD treatment is nastatinumab.

76. The method of claim 73, wherein the UC treatment is a composition comprising one or more bacterial species selected from the group consisting of: *Clostridium sacchariformis*, *Clostridium tenuifolium*, *Gemmelella formica*, *Ruminococcus tortifolia*, *Ostomyces viscerans*, *Collagenia fecalis*, *Bacillus sacchariformis*, *Collinus faecium*, *Vibrio vulnificus* species CAG: 241, *Lawsonia sacchariformis*, *Toxococcus butyricum*, *Clostridium* species CAG: 58, *Eubacterium* species CAG: 274, and *Bacillus warfarinia*, and optionally a pharmaceutically acceptable carrier.

77. The method of claim 73, wherein the CD treatment is a composition comprising one or more bacterial species selected from the group consisting of: *Ruminococcus ovalis* (Brutella ovalis), *Lawsonia ovalis* (unsaturated), *Roseobacterium inulinum*, *Dolphobicella formate*, *Eubacterium tumefaciens*, *Roseobacterium tumefaciens*, *Eubacterium tumefaciens*, *Anaerobic Bacillus hardruciformis*, *Vibrio vulnificus* species CAG: 241, *Vibrio vulnificus* species 57_20, *Lactobacillus rhamnosus*, *Eubacterium* species CAG: 274, and *Romebutzella ileum*, and optionally a pharmaceutically acceptable carrier.

78. The method as claimed in any of the preceding claims, wherein the first machine learning model and / or the second machine learning model is a generalized linear model.

79. A method for identifying a fecal sample with an altered inflammatory bowel disease (IBD) microbiome and / or an altered IBD subtype microbiome, wherein the IBD subtype is ulcerative colitis (UC) or Crohn's disease (CD), the method comprising the steps of: a) Determine the relative abundance of a first set of bacterial species, an optional level of fecal calprotectin, and an optional relative abundance of a second set of bacterial species in the fecal sample, wherein the first set of bacterial species and the second set of bacterial species each independently comprise one or more bacterial species selected from the group consisting of: Actinomycetes species oral taxa 181, Bacteroides fragilis, Escherichia coli, Lawsonia sucrase, Eubacterium species CAG: 274, Rosette of inulin, Rosette of the gut, Oval rumenococci (Oval Brutella), Formic acid-producing Dorl, Bileophilus warwickii, Clostridium truncatella, Clostridium sucrase, Formic acid-producing Gimeria, Visceral bacterium, Ruminococcus tortiformis, Clostridium spiralis, Gemmaella muscarinae, and Broutella hensenae; b) The relative abundance of each bacterial species in the first bacterial species set from the fecal sample is compared with the relative abundance of each bacterial species in the first bacterial species set in the first reference dataset, and optionally the fecal calprotectin level from the fecal sample is compared with the fecal calprotectin level in the first reference dataset using a first machine learning model to generate a risk score. as well as c) If the first risk score is higher than the first cutoff value, then the fecal sample is determined to have an altered IBD microbiome; d) If it is determined that the fecal sample has an altered IBD microbiome, then optionally a second machine learning model is used to compare the relative abundance of each bacterial species in the second bacterial species set from the fecal sample with the relative abundance of each bacterial species in the second bacterial species set in the second reference dataset to generate a second risk score. as well as e) Optionally, if the second risk score is higher than the second cutoff value, the fecal sample is determined to have an altered CD microbiome, or if the risk score is lower than or equal to the second cutoff value, the fecal sample is determined to have an altered UC microbiome.

80. The method of claim 79, wherein the first set of bacterial species comprises one or more bacterial species selected from the group consisting of: Actinomycetes species oral taxa 181, Escherichia coli, Lawsonia sucrase, Eubacterium species CAG: 274, Enterobacter rosenbergii, Ruminococcus ovalis (Broutella ovalis), Bileophilus warwickii, Clostridium truncatella, Clostridium sacchariformis, Gemmaformis, Osmotherium visceratum, Ruminococcus tortiformis, Clostridium spiralis, Gemmaella esculenta, and Brooutella henselae.

81. The method of claim 79 or claim 80, wherein the first set of bacterial species comprises two or more bacterial species, wherein the two or more bacterial species comprise a group of two bacterial species selected from the group consisting of: *Bacillus viscerans* and *Sclerotium mesenteriae*; *Bacillus viscerans* and *Clostridium velutipes*; *Bacillus viscerans* and *Brutella hensenoides*; *Bacillus viscerans* and *Escherichia coli*; *Bacillus viscerans* and *Actinomyces* species oral taxonomy group 181; *Sclerotium mesenteriae* and *Clostridium velutipes*; *Sclerotium mesenteriae* and *Brutella ovalis*; *Sclerotium mesenteriae* and *Brutella hensenoides*; *Sclerotium mesenteriae* and *Clostridium velutipes*; and *Sclerotium mesenteriae* and *Escherichia coli*.

82. The method of claim 79 or claim 80, wherein the first set of bacterial species comprises three or more bacterial species, wherein the three or more bacterial species comprise a group of three bacterial species selected from the group consisting of: *Ostomyces visceratus*, *Twincocephalus mesenteriae*, and *Clostridium velutipes*; *Ostomyces visceratus*, *Twincocephalus mesenteriae*, and *Brutella hensenoides*; *Ostomyces visceratus*, *Twincocephalus mesenteriae*, and *Clostridium spiralis*; *Twincocephalus mesenteriae*, *Clostridium velutipes*, and *Ruminococcus ovalis* (*Brutella ovalis*); *Ostomyces mesenteriae* Twin cocci, ovoid rumenococci (ovoid brouteria) and Hansen's brouteria; measles twin cocci, ovoid rumenococci (ovoid brouteria) and spiroclos; measles twin cocci, ovoid rumenococci (ovoid brouteria) and Escherichia coli; Clostridium tenuisae, ovoid rumenococci (ovoid brouteria) and Hansen's brouteria; Clostridium tenuisae, ovoid rumenococci (ovoid brouteria) and spiroclos; and Clostridium tenuisae, ovoid rumenococci (ovoid brouteria) and Escherichia coli.

83. The method of claim 79 or claim 80, wherein the first set of bacterial species comprises two or more bacterial species selected from the group consisting of: visceral bacilli, measles twin cocci, Clostridium tumefaciens, ovale broutococci (ovale broutococci), and broutococci hensenii.

84. The method of claim 79 or claim 80, wherein the first set of bacterial species comprises at least one bacterial species selected from the group consisting of: visceral bacterium, measles twin cocci, Clostridium tumefaciens and ovale rumenococci (ovale Brutella).

85. The method of claim 79 or claim 80, wherein the first set of bacterial species comprises at least one bacterial species selected from the group consisting of: visceral bacterium, measles twin cocci, Clostridium tumefaciens, ovale rumenococci (ovale Brutella) and bilirubinophilus.

86. The method of claim 79 or claim 80, wherein the first set of bacterial species comprises at least one bacterial species selected from the group consisting of: visceral bacterium, measles twin cocci, Clostridium truncatella, Brontë hanseniella, ovale rumenococci (ovale broutuli), and Clostridium spiralis.

87. The method of claim 79 or claim 80, wherein the first set of bacterial species comprises at least three bacterial species selected from the group consisting of: visceral bacterium, measles twin cocci, Clostridium truncatella, Brontë hanseniella, ovale rumenococci (ovale broutuli), and Clostridium spiralis.

88. The method of claim 79 or claim 80, wherein the first set of bacterial species comprises at least three bacterial species selected from the group consisting of: visceral bacterium, measles twin cocci, Clostridium truncatum, Broutella hensenae, Escherichia coli, and Clostridium spiralis.

89. The method of any one of claims 79-88, wherein the second set of bacterial species comprises two or more bacterial species, wherein the two or more bacterial species comprise two bacterial species selected from the group consisting of: *Brutella ovalis* and *Gimidae*; *Brutella ovalis* and *Roseobacterium rumenella*; *Brutella ovalis* and *Brutella Hansenii*; *Brutella ovalis* and *Roseobacterium inulinii*; *Brutella ovalis* and *Germococcus mesenteriae*; *Gimidae* and *Brutella Hansenii*; *Gimidae* and *Roseobacterium rumenella*; and *Gimidae* and *Germococcus mesenteriae*.

90. The method of any one of claims 79-88, wherein the second set of bacterial species comprises at least one bacterial species selected from the group consisting of: ovomycosis (Brutella ovalis), formic acid-producing Gemmaieux, Enterobacter rosenbergii, Broutella hensenae, Gemmaieux rubrum, and Bacteroides fragilis.

91. The method of any one of claims 79-88, wherein the second set of bacterial species comprises at least one bacterial species selected from the group consisting of: ovale rumenococci (ovale Brutella), formic acid germina, enterorosiformis, brutella hensenoides, measles twin cocci, unsaccharitrophoblastus, Bacteroides fragilis, and Escherichia coli.

92. The method of any one of claims 79-88, wherein the second set of bacterial species comprises at least one bacterial species selected from the group consisting of: ovomycosis (Broutella ovalis), formic acid-producing Gemmaieux, enterorosiformis, Brontotherium hensenoides, Gemella muscarinica, Lawsonia solani, Bacteroides fragilis, Escherichia coli, Clostridium truncatum, and Ostreae viscera.

93. The method of any of the preceding claims, wherein the first set of bacterial species comprises four or more bacterial species selected from the group consisting of: Actinomycetes oral taxa 181, Escherichia coli, Lawsonia solani, Eubacterium species CAG: 274, Enterobacter rosenbergii, Ruminococcus ovalis (Broutella ovalis), Biliophilus warwickii, Clostridium truncatella, Clostridium sacchariformis, Formicoblastic Gemmaieux, Ostomyces viscerans, Ruminococcus tortiformis, Clostridium spiralis, Gemella mesenteriae, and Brontë hansenii, and the second set of bacterial species comprises four or more bacterial species selected from the group consisting of: Actinomycetes oral taxa 181, Bacteroides fragilis, Escherichia coli, Lawsonia solani, Eubacterium species CAG:

274.

274. Inulin-eating Roselle, Enteric Roselle, Oval-shaped Ruminococcus (Oval Broutella), Formic acid-producing Dorl, Bileophilus warwickii, Clostridium truncatella, Clostridium sacchariformis, Formic acid-producing Gemma, Visceral Odorbacterium, Twisted Ruminococcus, Clostridium spiralis, Measles twin cocci, and Broutella hensenoides.

94. The method as claimed in any of the preceding claims, wherein the first set of bacterial species comprises substantially the following: The oral taxonomy of Actinomycetes species 181, *Escherichia coli*, *Lawsonia unagalactiae*, *Eubacterium* species CAG: 274, *Roseidonella enterica*, *Broutella ovalis*, *Biliobacterium warfarinii*, *Clostridium truncatella*, *Clostridium sacchariformis*, *Gemmelella formica*, *Ostomyces viscerans*, *Ruminococcus tortifolia*, *Clostridium spiralis*, *Twincocephala spp.*, and *Broutella henselae*, and the second bacterial species set is basically composed of the following: oral taxonomy of Actinomycetes species 181, *Bacteroides fragilis*, *Escherichia coli*, *Lawsonia unagalactiae*, *Eubacterium* species CAG:

274. Inulin-eating Roselle, Enteric Roselle, Oval-shaped Ruminococcus (Oval Broutella), Formic acid-producing Dorl, Bileophilus warwickii, Clostridium truncatella, Clostridium sacchariformis, Formic acid-producing Gemma, Visceral Odorbacterium, Twisted Ruminococcus, Clostridium spiralis, Measles twin cocci, and Broutella hensenoides.

95. The method as described in any of the preceding claims, wherein The first reference dataset was obtained by determining the relative abundance of each bacterial species in the first bacterial species set and, optionally, the level of fecal calprotectin in fecal samples from a first subject group comprising patients diagnosed with IBD and individuals without IBD in a reference cohort; and The second reference dataset was obtained by determining the relative abundance of each bacterial species in the second bacterial species set in fecal samples from a second subject group, which included patients diagnosed with UC and patients diagnosed with CD.

96. The method of claim 95, wherein the subjects in the second subject group are selected from patients in the first subject group who have been diagnosed with IBD.

97. The method as described in any of the preceding claims, wherein the first machine learning model is constructed by: i) The first reference dataset is obtained by determining the relative abundance of each bacterial species in the first bacterial species set and the optional fecal calprotectin level in fecal samples from the first subject group, which includes patients diagnosed with IBD and individuals in the reference cohort who do not have IBD. ii) Divide the first reference dataset into a first training dataset and a first test dataset; iii) Divide the first training dataset into a first training subset and a first validation subset; iv) Train the first candidate model using the first training subset and evaluate the performance of the first candidate model using the first validation subset; v) Optionally, the first candidate model can be tuned by adjusting the hyperparameters in the first candidate model; vi) Repeat steps (iv) and optional steps (v) to generate a first candidate model set and select the best first candidate model from the first candidate model set; vii) Evaluate the performance of the best first candidate model using the first test dataset; viii) Optionally, using the hyperparameter set corresponding to the best first candidate model, a first final model is trained with the first reference dataset to produce the first final model; as well as ix) Use the best first candidate model from step (vii), or if the first final model was trained using step (viii), use the first final model from step (viii) as the first machine learning model for the comparison step in step (b) to generate the first risk score.

98. The method as described in any of the preceding claims, wherein the second machine learning model is constructed by: i. The second reference dataset is obtained by determining the relative abundance of each bacterial species in the second bacterial species set from fecal samples from the second subject group, which includes patients diagnosed with UC and patients diagnosed with CD; ii. Divide the second reference dataset into a second training dataset and a second test dataset; iii. Divide the second training dataset into a second training subset and a second validation subset; iv. Train the second candidate model using the second training subset and evaluate the performance of the second candidate model using the second validation subset; v. Optionally, the second candidate model can be tuned by adjusting the hyperparameters in the second candidate model; vi. Repeat steps (iv) and optional steps (v) to generate a second set of candidate models, and select the best second candidate model from the second set of candidate models; vii. Evaluate the performance of the best second candidate model using the second test dataset; viii. Using the hyperparameter set corresponding to the best second candidate model, train the second final model with the second reference dataset to produce the second final model; and ix. Use the best second candidate model from step (vii), or if the second final model was trained using step (viii), use the second final model from step (viii) as the second machine learning model for the comparison step in step (d) to generate the second risk score.

99. The method as claimed in any of the preceding claims, wherein the first machine learning model is a first random forest model, and wherein the first risk score in step (b) is generated by the following steps: (b1) Using the relative abundance of the first bacterial species set in the first reference dataset and the optional fecal calprotectin level, generate a set of decision trees using the first random forest model; and (b2) Run the relative abundance of the first bacterial species set from the fecal sample and the optional fecal calprotectin level along the decision tree set to generate the first risk score.

100. The method as claimed in any of the preceding claims, wherein the second machine learning model is a second random forest model, and wherein the second risk score in step (d) is generated by the following steps: (d1) Using the relative abundance of the second bacterial species set in the second reference dataset, generate a set of decision trees using the second random forest model; and (d2) Run the relative abundance of the second set of bacterial species from the fecal sample along the decision tree set to generate the second risk score.

101. The method of claim 97, wherein the best first candidate model is the first candidate model with the highest performance in the first candidate model set.

102. The method of claim 98, wherein the best second candidate model is the second candidate model with the highest performance in the second candidate model set.

103. The method of any of the preceding claims, wherein the performance of the first candidate model and / or the second candidate model is evaluated by the area under the curve (AUC) of the receiver operating characteristic (ROC) curve of the first candidate model and / or the second candidate model.

104. The method as claimed in any of the preceding claims, wherein the first cutoff value and / or the second cutoff value are in the range of 0.4 to 0.

75.

105. The method as claimed in any of the preceding claims, wherein the first cutoff value is determined by the following steps: (c1) Calculate the Yoden Index using the first risk score data obtained from the first reference dataset; and (c2) Determine the first cutoff value based on the Yoden index.

106. The method as claimed in any of the preceding claims, wherein the second cutoff value is determined by the following steps: (e1) Calculate the Yoden Index using the second risk score data obtained from the second reference dataset; and (e2) Determine the second cutoff value based on the Yoden index.

107. The method as claimed in any of the preceding claims, wherein the IBD is an active IBD or an inactive IBD, the UC is an active UC or an inactive UC, and the CD is an active CD or an inactive CD.

108. The method of any of the preceding claims, wherein the relative abundance of the first bacterial species set and the relative abundance of the second bacterial species set are determined by metagenomic sequencing and / or PCR.

109. The method of claim 108, wherein the PCR is a digital droplet PCR (ddPCR) or quantitative PCR (qPCR) performed using one or more primer sets and optionally one or more fluorophore-quencher probes, wherein each primer set is used to amplify a target nucleic acid sequence of a bacterial species belonging to the first bacterial species set or the second bacterial species set to produce an amplification product, and each fluorophore-quencher probe is used to detect the amplification product of each primer set.

110. The method of claim 109, wherein the target nucleic acid sequence amplified by each primer set comprises a portion of a nucleotide sequence selected from the group consisting of: SEQ ID NO: 76, SEQ ID NO: 77, SEQ ID NO: 78, SEQ ID NO: 80, SEQ ID NO: 81, SEQ ID NO: 82, SEQ ID NO: 85, SEQ ID NO: 89, SEQ ID NO: 92, SEQ ID NO: 137, SEQ ID NO: 139, SEQ ID NO: 141, SEQ ID NO: 143, SEQ ID NO: 145, SEQ ID NO: 147, SEQ ID NO: 160, SEQ ID NO: 162, and SEQ ID NO:

164.

111. The method of claim 109, wherein the target nucleic acid sequence amplified by each primer set comprises a nucleotide sequence selected from the group consisting of: SEQ ID NO: 58, SEQ ID NO: 59, SEQ ID NO: 60, SEQ ID NO: 62, SEQ ID NO: 63, SEQ ID NO: 64, SEQ ID NO: 67, SEQ ID NO: 71, SEQ ID NO: 74, SEQ ID NO: 136, SEQ ID NO: 138, SEQ ID NO: 140, SEQ ID NO: 142, SEQ ID NO: 144, SEQ ID NO: 146, SEQ ID NO: 159, SEQ ID NO: 161 and SEQ ID NO:

163.

112. The method of claim 109, wherein the one or more primer sets comprise the forward primer of SEQ ID NO: 1 and the reverse primer of SEQ ID NO: 2; the forward primer of SEQ ID NO: 3 and the reverse primer of SEQ ID NO: 4; the forward primer of SEQ ID NO: 5 and the reverse primer of SEQ ID NO: 6; the forward primer of SEQ ID NO: 9 and the reverse primer of SEQ ID NO: 10; the forward primer of SEQ ID NO: 11 and the reverse primer of SEQ ID NO: 12; the forward primer of SEQ ID NO: 13 and the reverse primer of SEQ ID NO: 14; the forward primer of SEQ ID NO: 19 and the reverse primer of SEQ ID NO: 20; the forward primer of SEQ ID NO: 33 and the reverse primer of SEQ ID NO: 34; the forward primer of SEQ ID NO: 94 and the reverse primer of SEQ ID NO: 95; SEQ ID NO: The forward primer of SEQ ID NO: 96 and the reverse primer of SEQ ID NO: 97; the forward primer of SEQ ID NO: 98 and the reverse primer of SEQ ID NO: 99; the forward primer of SEQ ID NO: 100 and the reverse primer of SEQ ID NO: 101; the forward primer of SEQ ID NO: 102 and the reverse primer of SEQ ID NO: 103; the forward primer of SEQ ID NO: 104 and the reverse primer of SEQ ID NO: 105; the forward primer of SEQ ID NO: 106 and the reverse primer of SEQ ID NO: 107; the forward primer of SEQ ID NO: 118 and the reverse primer of SEQ ID NO: 119; and / or the forward primer of SEQ ID NO: 120 and the reverse primer of SEQ ID NO: 121; and each fluorophore-quencher probe comprises a probe sequence selected from the group consisting of: SEQ ID NO: 39, SEQ ID NO: 40, SEQ ID NO: 96, SEQ ID NO: 97, SEQ ID NO: 98, SEQ ID NO: 99, SEQ ID NO: 99, SEQ ID NO: 100, SEQ ID NO: 101, SEQ ID NO: 102, SEQ ID NO: 103, SEQ ID NO: 104, SEQ ID NO: 105, SEQ ID NO: 106, SEQ ID NO: 107, SEQ ID NO: 118, SEQ ID NO: 119, and / or the forward primer of SEQ ID NO: 120 and the reverse primer of SEQ ID NO: 121; and each fluorophore-quencher probe comprises a probe sequence selected from the group consisting of: SEQ ID NO: 39, SEQ ID NO: 40, SEQ ID NO: 99, SEQ ID NO: 99, SEQ ID NO: 99, SEQ ID NO: ID NO: 41, SEQ ID NO: 43, SEQ ID NO: 44, SEQ ID NO: 45, SEQ ID NO: 48, SEQ ID NO: 52, SEQ ID NO: 55, SEQ ID NO: 122, SEQ ID NO: 123, SEQ ID NO: 124, SEQ ID NO: 125, SEQ ID NO: 126, SEQ ID NO: 127, SEQ ID NO: 128, SEQ ID NO: 134 and SEQ ID NO:

135.

113. A method for identifying a fecal sample with an altered IBD subtype microbiome, wherein the IBD subtype is ulcerative colitis (UC) or Crohn's disease (CD), the method comprising the steps of: a) Determine whether the fecal sample has an altered IBD microbiome; b) If it is determined that the fecal sample has an altered IBD microbiome, then determine the relative abundance of a second set of bacterial species in the fecal sample from the individual, wherein the second set of bacterial species comprises one or more bacterial species selected from the group consisting of: Actinomycete species oral taxa 181, Bacteroides fragilis, Escherichia coli, Lawsonia sucrase, Eubacterium species CAG: 274, Rosette of inulin, Rosette of the gut, Oval rumenococcus (Oval Brutella), Formic acid-producing Dorl, Bileophilus warwickii, Clostridium truncatella, Clostridium sucrase, Formic acid-producing G. formic acid, Osmophilus viscera, Ruminococcus tortiformis, Clostridium spiralis, Gemmaella muscarinae, and Broutella hensenae; c) Using a second machine learning model, the relative abundance of each bacterial species in the second bacterial species set from the fecal sample is compared with the relative abundance of each bacterial species in the second bacterial species set in the second reference dataset to generate a second risk score; as well as d) If the second risk score is higher than the second cutoff value, the fecal sample is determined to have an altered CD microbiome; or if the risk score is lower than or equal to the second cutoff value, the fecal sample is determined to have an altered UC microbiome.

114. The method of claim 113, wherein the determining step in step (a) further comprises the following steps: (a1) Determine the relative abundance of a first set of bacterial species and, optionally, the level of fecal calprotectin in the fecal sample, wherein the first set of bacterial species comprises one or more bacterial species selected from the group consisting of: Actinomycetes species oral taxa 181, Bacteroides fragilis, Escherichia coli, Lawsonia sucrase, Eubacterium species CAG: 274, Rosette of inulin, Rosette of the gut, Oval rumenococci (Oval Brutella), Formic acid-producing Dorl, Bileophilus warwickii, Clostridium truncatella, Clostridium sucrase, Formic acid-producing G. formic acid, Osmophilus viscera, Ruminococcus tortiformis, Clostridium spiralis, Gemmaella muscarinae, and Broutella hensenae; (a2) The relative abundance of each bacterial species in the first bacterial species set from the fecal sample is compared with the relative abundance of each bacterial species in the first bacterial species set in the first reference dataset, and optionally the fecal calprotectin level from the fecal sample is compared with the fecal calprotectin level in the first reference dataset using a first machine learning model to generate a first risk score. (a3) If the first risk score is higher than the first cutoff value, the fecal sample is determined to have an altered IBD microbiome.

115. The method of claim 113 or claim 114, wherein the second set of bacterial species comprises two or more bacterial species, wherein the two or more bacterial species comprise two bacterial species selected from the group consisting of: *Brutella ovalis* and *Gimidae*; *Brutella ovalis* and *Roseobacterium rumenella*; *Brutella ovalis* and *Brutella Hansenii*; *Brutella ovalis* and *Roseobacterium inulinii*; *Brutella ovalis* and *Germococcus mesenteriae*; *Gimidae* and *Brutella Hansenii*; *Gimidae* and *Roseobacterium rumenella*; and *Gimidae* and *Germococcus mesenteriae*.

116. The method of claim 113 or claim 114, wherein the second set of bacterial species comprises at least one bacterial species selected from the group consisting of: ovomycosis (Brutella ovalis), formic acid-producing Gemmaieux, Enterobacter rosenbergii, Broutella hensenae, Gemmaieux rubrum, and Bacteroides fragilis.

117. The method of claim 113 or claim 114, wherein the second set of bacterial species comprises at least one bacterial species selected from the group consisting of: ovomycosis (Brutella ovalis), formic acid-producing Gemmaieux, Enterobacter rosenbergii, Brautella hensenae, Gemmaieux malariae, Lawsonia sacchariformis, Bacteroides fragilis, and Escherichia coli.

118. The method of claim 113 or claim 114, wherein the second set of bacterial species comprises at least one bacterial species selected from the group consisting of: ovomycosis (Broutella ovalis), formic acid-producing Gemmaieux, enteric Rosebes, Hansenella broutella, measles twin cocci, sugar-insoluble Lawsonia, Bacteroides fragilis, Escherichia coli, Clostridium truncatum, and visceral bacterium.

119. The method of any of the preceding claims, wherein the second set of bacterial species comprises four or more bacterial species selected from the group consisting of: Actinomycetes species oral taxa 181, Bacteroides fragilis, Escherichia coli, Lawsonia sucrase, Eubacterium species CAG: 274, Rosette of inulin, Rosette of the gut, Oval rumenococcus (Oval Brutella), Formic acid-producing Dorl, Bileophilus warwickii, Clostridium truncatella, Clostridium sucrase, Formic acid-producing G. formic acid, Gemmaella formica, Visceral bacterium, Ruminococcus tortiformis, Clostridium spiralis, Gemmaella muscarinae, and Broutella hensenii.

120. The method as described in any of the preceding claims, wherein the second set of bacterial species comprises substantially the following: Actinomycetes species oral taxonomy group 181, Bacteroides fragilis, Escherichia coli, Lawsonia sacchariformis, Eubacterium species CAG: 274, Rosette sacchariformis, Rosette sacchariformis, Brontë ovale, Formico-producing Dorl, Bileophilus warwickii, Clostridium truncatella, Clostridium sacchariformis, Formico-producing Clostridium, Gemite bacilli, Visceral bacterium, Ruminococcus tortiformis, Clostridium spiralis, Gemite bacilli, and Brontë hansenii.

121. The method as described in any of the preceding claims, wherein The second reference dataset was obtained by determining the relative abundance of each bacterial species in the second bacterial species set in fecal samples from a second subject group, which included patients diagnosed with UC and patients diagnosed with CD.

122. The method of claim 114, wherein the first reference dataset is obtained by determining the relative abundance of each bacterial species in the first bacterial species set and, optionally, the level of fecal calprotectin in fecal samples from a first subject group, the first subject group comprising patients diagnosed with IBD and individuals without IBD in a reference cohort; and wherein the subjects in the second subject group are selected from the patients diagnosed with IBD in the first subject group.

123. The method as described in any of the preceding claims, wherein the second machine learning model is constructed by: i. The second reference dataset is obtained by determining the relative abundance of each bacterial species in the second bacterial species set from fecal samples from the second subject group, which includes patients diagnosed with UC and patients diagnosed with CD; ii. Divide the second reference dataset into a second training dataset and a second test dataset; iii. Divide the second training dataset into a second training subset and a second validation subset; iv. Train the second candidate model using the second training subset and evaluate the performance of the second candidate model using the second validation subset; v. Optionally, the second candidate model can be tuned by adjusting the hyperparameters in the second candidate model; vi. Repeat steps (iv) and optional steps (v) to generate a second set of candidate models, and select the best second candidate model from the second set of candidate models; vii. Evaluate the performance of the best second candidate model using the second test dataset; viii. Using the hyperparameter set corresponding to the best second candidate model, train the second final model with the second reference dataset to produce the second final model; and ix. Use the best second candidate model from step (vii), or if the second final model was trained using step (viii), use the second final model from step (viii) as the second machine learning model for the comparison step in step (d) to generate the second risk score.

124. The method as claimed in any of the preceding claims, wherein the second machine learning model is a second random forest model, and wherein the second risk score in step (c) is generated by the following steps: (c1) Using the relative abundance of the second bacterial species set in the second reference dataset, generate a set of decision trees using the second random forest model; and (c2) Run the relative abundance of the second set of bacterial species from the individuals along the decision tree set to generate the second risk score.

125. The method of claim 123, wherein the best second candidate model is the second candidate model with the highest performance in the second candidate model set.

126. The method of any one of claims 123-125, wherein the performance of the second candidate model is evaluated by the area under the curve (AUC) of the receiver operating characteristic (ROC) curve of the second candidate model.

127. The method as claimed in any of the preceding claims, wherein the second cutoff value is in the range of 0.4 to 0.

75.

128. The method as claimed in any of the preceding claims, wherein the second cutoff value is determined by the following steps: (e1) Calculate the Yoden Index using the second risk score data obtained from the second reference dataset; and (e2) Determine the second cutoff value based on the Yoden index.

129. The method as claimed in any of the preceding claims, wherein the IBD is an active IBD or an inactive IBD, the UC is an active UC or an inactive UC, and the CD is an active CD or an inactive CD.

130. The method of any of the preceding claims, wherein the relative abundance of the first bacterial species set and / or the second bacterial species set is determined by metagenomic sequencing and / or PCR.

131. The method of claim 130, wherein the PCR is a digital droplet PCR (ddPCR) or quantitative PCR (qPCR) performed using one or more primer sets and optionally one or more fluorophore-quencher probes, wherein each primer set is used to amplify a target nucleic acid sequence of a bacterial species belonging to the first bacterial species set or the second bacterial species set to produce an amplification product, and each fluorophore-quencher probe is used to detect the amplification product of each primer set.

132. The method of claim 131, wherein the target nucleic acid sequence amplified by each primer set comprises a portion of a nucleotide sequence selected from the group consisting of: SEQ ID NO: 76, SEQ ID NO: 77, SEQ ID NO: 78, SEQ ID NO: 80, SEQ ID NO: 81, SEQ ID NO: 82, SEQ ID NO: 85, SEQ ID NO: 89, SEQ ID NO: 92, SEQ ID NO: 137, SEQ ID NO: 139, SEQ ID NO: 141, SEQ ID NO: 143, SEQ ID NO: 145, SEQ ID NO: 147, SEQ ID NO: 160, SEQ ID NO: 162, and SEQ ID NO:

164.

133. The method of claim 131, wherein the target nucleic acid sequence amplified by each primer set comprises a nucleotide sequence selected from the group consisting of: SEQ ID NO: 58, SEQ ID NO: 59, SEQ ID NO: 60, SEQ ID NO: 62, SEQ ID NO: 63, SEQ ID NO: 64, SEQ ID NO: 67, SEQ ID NO: 71, SEQ ID NO: 74, SEQ ID NO: 136, SEQ ID NO: 138, SEQ ID NO: 140, SEQ ID NO: 142, SEQ ID NO: 144, SEQ ID NO: 146, SEQ ID NO: 159, SEQ ID NO: 161 and SEQ ID NO:

163.

134. The method of claim 131, wherein the one or more primer sets comprise the forward primer of SEQ ID NO: 1 and the reverse primer of SEQ ID NO: 2; the forward primer of SEQ ID NO: 3 and the reverse primer of SEQ ID NO: 4; the forward primer of SEQ ID NO: 5 and the reverse primer of SEQ ID NO: 6; the forward primer of SEQ ID NO: 9 and the reverse primer of SEQ ID NO: 10; the forward primer of SEQ ID NO: 11 and the reverse primer of SEQ ID NO: 12; the forward primer of SEQ ID NO: 13 and the reverse primer of SEQ ID NO: 14; the forward primer of SEQ ID NO: 19 and the reverse primer of SEQ ID NO: 20; the forward primer of SEQ ID NO: 33 and the reverse primer of SEQ ID NO: 34; the forward primer of SEQ ID NO: 94 and the reverse primer of SEQ ID NO: 95; SEQ ID NO: The forward primer of SEQ ID NO: 96 and the reverse primer of SEQ ID NO: 97; the forward primer of SEQ ID NO: 98 and the reverse primer of SEQ ID NO: 99; the forward primer of SEQ ID NO: 100 and the reverse primer of SEQ ID NO: 101; the forward primer of SEQ ID NO: 102 and the reverse primer of SEQ ID NO: 103; the forward primer of SEQ ID NO: 104 and the reverse primer of SEQ ID NO: 105; the forward primer of SEQ ID NO: 106 and the reverse primer of SEQ ID NO: 107; the forward primer of SEQ ID NO: 118 and the reverse primer of SEQ ID NO: 119; and / or the forward primer of SEQ ID NO: 120 and the reverse primer of SEQ ID NO: 121; and each fluorophore-quencher probe comprises a probe sequence selected from the group consisting of: SEQ ID NO: 39, SEQ ID NO: 40, SEQ ID NO: 96, SEQ ID NO: 97, SEQ ID NO: 98, SEQ ID NO: 99, SEQ ID NO: 99, SEQ ID NO: 100, SEQ ID NO: 101, SEQ ID NO: 102, SEQ ID NO: 103, SEQ ID NO: 104, SEQ ID NO: 105, SEQ ID NO: 106, SEQ ID NO: 107, SEQ ID NO: 118, SEQ ID NO: 119, and / or the forward primer of SEQ ID NO: 120 and the reverse primer of SEQ ID NO: 121; and each fluorophore-quencher probe comprises a probe sequence selected from the group consisting of: SEQ ID NO: 39, SEQ ID NO: 40, SEQ ID NO: 99, SEQ ID NO: 99, SEQ ID NO: 99, SEQ ID NO: ID NO: 41, SEQ ID NO: 43, SEQ ID NO: 44, SEQ ID NO: 45, SEQ ID NO: 48, SEQ ID NO: 52, SEQ ID NO: 55, SEQ ID NO: 122, SEQ ID NO: 123, SEQ ID NO: 124, SEQ ID NO: 125, SEQ ID NO: 126, SEQ ID NO: 127, SEQ ID NO: 128, SEQ ID NO: 134 and SEQ ID NO:

135.

135. A kit for determining the risk or diagnosing inflammatory bowel disease (IBD) and / or IBD subtypes, wherein the IBD subtype is ulcerative colitis (UC) or Crohn's disease (CD) in an individual, the kit comprising reagents for detecting a first set of bacterial species and optionally reagents for detecting a second set of bacterial species, wherein the first set of bacterial species and the second set of bacterial species each independently comprise one or more bacterial species selected from the group consisting of: Actinomycetes species oral taxa 181, Bacteroides fragilis, Escherichia coli, Lawsonia sucrase, Eubacterium species CAG:274, Rosette of inulin, Rosette of the gut, Oval rumenococci (Oval Brutella), Formic acid-producing Dorl, Bileophilus warwickii, Clostridium truncatella, Clostridium sucrase, Formic acid-producing Gemmaella, Visceral Odorobacter, Ruminococcus tortiformis, Clostridium spiralis, Gemmaella muscarinae, and Broutella hensenii.

136. The kit of claim 135, wherein the first set of bacterial species comprises one or more bacterial species selected from the group consisting of: Actinomycetes species oral taxa 181, Escherichia coli, Lawsonia sucrase, Eubacterium species CAG: 274, Enterobacter rosenbergii, Ruminococcus ovalis (Broutella ovalis), Bileophilus warwickii, Clostridium truncatella, Clostridium sacchariformis, Gemmaformis, Osmotherium visceratum, Ruminococcus tortiformis, Clostridium spiralis, Gemmaella esculenta, and Brooutella henselae.

137. The kit of claim 135 or claim 136, wherein the first set of bacterial species comprises two or more bacterial species, wherein the two or more bacterial species comprise two bacterial species selected from the group consisting of: *Bacillus viscerans* and *Sclerotium mesenteriae*; *Bacillus viscerans* and *Clostridium velutipes*; *Bacillus viscerans* and *Brutella hensenoides*; *Bacillus viscerans* and *Escherichia coli*; *Bacillus viscerans* and *Actinomyces* species oral taxonomy group 181; *Sclerotium mesenteriae* and *Clostridium velutipes*; *Sclerotium mesenteriae* and *Ruminococcus ovalis* (*Brutella ovalis*); *Sclerotium mesenteriae* and *Brutella hensenoides*; *Sclerotium mesenteriae* and *Clostridium velutipes*; and *Sclerotium mesenteriae* and *Escherichia coli*.

138. The kit of claim 135 or claim 136, wherein the first set of bacterial species comprises three or more bacterial species, wherein the three or more bacterial species comprise a group of three bacterial species selected from the group consisting of: *Ostomyces visceratus*, *Twincocephalus mesenteriae*, and *Clostridium velutipes*; *Ostomyces visceratus*, *Twincocephalus mesenteriae*, and *Brutella hensenoides*; *Ostomyces visceratus*, *Twincocephalus mesenteriae*, and *Clostridium spiralis*; *Twincocephalus mesenteriae*, *Clostridium velutipes*, and *Ruminococcus ovalis* (Brutella ovalis). Measles twin cocci, ovoid rumenococci (ovoid brouté) and Hansen's brouté; measles twin cocci, ovoid rumenococci (ovoid brouté) and Clostridium spiralis; measles twin cocci, ovoid rumenococci (ovoid brouté) and Escherichia coli; Clostridium tenuisae, ovoid rumenococci (ovoid brouté) and Hansen's brouté; Clostridium tenuisae, ovoid rumenococci (ovoid brouté) and Clostridium spiralis; and Clostridium tenuisae, ovoid rumenococci (ovoid brouté) and Escherichia coli.

139. The kit of claim 135 or claim 136, wherein the first set of bacterial species comprises two or more bacterial species selected from the group consisting of: visceral bacterium, measles twin cocci, Clostridium tumefaciens, ovale rumenococci (ovale Brutella) and brutella hensenii.

140. The kit of claim 135 or claim 136, wherein the first set of bacterial species comprises at least one bacterial species selected from the group consisting of: visceral bacterium, measles twin cocci, Clostridium tumefaciens and ovoid rumenococci (ovoid broutella).

141. The kit of claim 135 or claim 136, wherein the first set of bacterial species comprises at least one bacterial species selected from the group consisting of: visceral bacterium, measles twin cocci, Clostridium tumefaciens, ovale rumenococci (ovale Brutella) and bilirubinophilus.

142. The kit of claim 135 or claim 136, wherein the first set of bacterial species comprises at least one bacterial species selected from the group consisting of: visceral bacterium, measles twin cocci, Clostridium truncatella, Brontë hanseniella, ovale rumenococci (ovale broutuli), and Clostridium spiralis.

143. The kit of claim 135 or claim 136, wherein the first set of bacterial species comprises at least three bacterial species selected from the group consisting of: visceral bacterium, measles twin cocci, Clostridium truncatella, Brontë hanseniella, ovale rumenococci (ovale broutuli), and Clostridium spiralis.

144. The kit of claim 135 or claim 136, wherein the first set of bacterial species comprises at least three bacterial species selected from the group consisting of: visceral bacterium, Geminidella mesenteriae, Clostridium truncatella, Broutella hensenii, Escherichia coli, and Clostridium spiralis.

145. The kit according to any one of claims 135-144, wherein the second set of bacterial species comprises two or more bacterial species, wherein the two or more bacterial species comprise two bacterial species selected from the group consisting of: *Brutella ovalis* and *G. formic acid*; *Brutella ovalis* and *R. rostrella*; *Brutella ovalis* and *Brutella hensenii*; *Brutella ovalis* and *R. rostrella*; *Brutella ovalis* and *G. rostrella*; *G. formic acid* and *Brutella hensenii*; *G. formic acid* and *R. rostrella*; and *G. formic acid* and *G. rostrella*.

146. The kit according to any one of claims 135-144, wherein the second set of bacterial species comprises at least one bacterial species selected from the group consisting of: ovomycosis (Brutella ovalis), formic acid-producing Gemmaietta var. formica, Rosella ovalis, Broutella hensenii, Gemella mesenteriae, and Bacteroides fragilis.

147. The kit according to any one of claims 135-144, wherein the second set of bacterial species comprises at least one bacterial species selected from the group consisting of: ovomycosis (Brutella ovalis), formic acid-producing Gemmaieux, Enterobacter rosenbergii, Brautella hensenae, Gemmaieux var. measlese, Lawsonia fragilis, Bacteroides fragilis, and Escherichia coli.

148. The kit according to any one of claims 135-144, wherein the second set of bacterial species comprises at least one bacterial species selected from the group consisting of: ovomycosis (Brutella ovalis), formic acid-producing Gemmaieux, enteric Rosebes, Brautella hensenae, Gemmaieux var. measles, Lawsonia solani, Bacteroides fragilis, Escherichia coli, Clostridium truncatum, and Osmophilus viscera.

149. The kit according to any one of claims 135-144, wherein the first bacterial species set comprises four or more bacterial species selected from the group consisting of: Actinomycetes oral taxa 181, Escherichia coli, Lawsonia solani, Eubacterium species CAG: 274, Enterobacter rosenbergii, Ruminococcus ovalis (Broutella ovalis), Biliophilus warwickii, Clostridium truncatella, Clostridium sacchariformis, Formic acid-producing Gemmaieux, Ostomyces viscerans, Ruminococcus tortiformis, Clostridium spiralis, Gemella mesenteriae, and Brontë hansenii, and the second bacterial species set comprises four or more bacterial species selected from the group consisting of: Actinomycetes oral taxa 181, Bacteroides fragilis, Escherichia coli, Lawsonia solani, Eubacterium species CAG:

274.

274. Inulin-eating Roselle, Enteric Roselle, Oval-shaped Ruminococcus (Oval Broutella), Formic acid-producing Dorl, Bileophilus warwickii, Clostridium truncatella, Clostridium sacchariformis, Formic acid-producing Gemma, Visceral Odorbacterium, Twisted Ruminococcus, Clostridium spiralis, Measles twin cocci, and Broutella hensenoides.

150. The kit as claimed in any of the preceding claims, wherein the first set of bacterial species comprises substantially the following: The oral taxonomy of Actinomycetes species 181, *Escherichia coli*, *Lawsonia unagalactiae*, *Eubacterium* species CAG: 274, *Roseidonella enterica*, *Broutella ovalis*, *Biliobacterium warfarinii*, *Clostridium truncatella*, *Clostridium sacchariformis*, *Gemmelella formica*, *Ostomyces viscerans*, *Ruminococcus tortifolia*, *Clostridium spiralis*, *Twincocephala spp.*, and *Broutella henselae*, and the second bacterial species set is basically composed of the following: oral taxonomy of Actinomycetes species 181, *Bacteroides fragilis*, *Escherichia coli*, *Lawsonia unagalactiae*, *Eubacterium* species CAG:

274. Inulin-eating Roselle, Enteric Roselle, Oval-shaped Ruminococcus (Oval Broutella), Formic acid-producing Dorl, Bileophilus warwickii, Clostridium truncatella, Clostridium sacchariformis, Formic acid-producing Gemma, Visceral Odorbacterium, Twisted Ruminococcus, Clostridium spiralis, Measles twin cocci, and Broutella hensenoides.

151. The kit according to any one of the preceding claims, wherein the reagent comprises one or more primer sets and optionally one or more fluorophore-quencher probes, wherein each primer set is used to amplify a target nucleic acid sequence of a bacterial species belonging to the first bacterial species set or the second bacterial species set via PCR to produce an amplification product, and each fluorophore-quencher probe is used to detect the amplification product of each primer set.

152. The kit of claim 151, wherein the target nucleic acid sequence amplified by each primer set comprises a portion of a nucleotide sequence selected from the group consisting of: SEQ ID NO: 76, SEQ ID NO: 77, SEQ ID NO: 78, SEQ ID NO: 80, SEQ ID NO: 81, SEQ ID NO: 82, SEQ ID NO: 85, SEQ ID NO: 89, SEQ ID NO: 92, SEQ ID NO: 137, SEQ ID NO: 139, SEQ ID NO: 141, SEQ ID NO: 143, SEQ ID NO: 145, SEQ ID NO: 147, SEQ ID NO: 160, SEQ ID NO: 162, and SEQ ID NO:

164.

153. The kit of claim 151, wherein the target nucleic acid sequence amplified by each primer set comprises a nucleotide sequence selected from the group consisting of: SEQ ID NO: 58, SEQ ID NO: 59, SEQ ID NO: 60, SEQ ID NO: 62, SEQ ID NO: 63, SEQ ID NO: 64, SEQ ID NO: 67, SEQ ID NO: 71, SEQ ID NO: 74, SEQ ID NO: 136, SEQ ID NO: 138, SEQ ID NO: 140, SEQ ID NO: 142, SEQ ID NO: 144, SEQ ID NO: 146, SEQ ID NO: 159, SEQ ID NO: 161 and SEQ ID NO:

163.

154. The kit of claim 151, wherein the one or more primer sets comprise the forward primer of SEQ ID NO: 1 and the reverse primer of SEQ ID NO: 2; the forward primer of SEQ ID NO: 3 and the reverse primer of SEQ ID NO: 4; the forward primer of SEQ ID NO: 5 and the reverse primer of SEQ ID NO: 6; the forward primer of SEQ ID NO: 9 and the reverse primer of SEQ ID NO: 10; the forward primer of SEQ ID NO: 11 and the reverse primer of SEQ ID NO: 12; the forward primer of SEQ ID NO: 13 and the reverse primer of SEQ ID NO: 14; the forward primer of SEQ ID NO: 19 and the reverse primer of SEQ ID NO: 20; the forward primer of SEQ ID NO: 33 and the reverse primer of SEQ ID NO: 34; the forward primer of SEQ ID NO: 94 and the reverse primer of SEQ ID NO: 95; SEQ ID NO: The forward primer of SEQ ID NO: 96 and the reverse primer of SEQ ID NO: 97; the forward primer of SEQ ID NO: 98 and the reverse primer of SEQ ID NO: 99; the forward primer of SEQ ID NO: 100 and the reverse primer of SEQ ID NO: 101; the forward primer of SEQ ID NO: 102 and the reverse primer of SEQ ID NO: 103; the forward primer of SEQ ID NO: 104 and the reverse primer of SEQ ID NO: 105; the forward primer of SEQ ID NO: 106 and the reverse primer of SEQ ID NO: 107; the forward primer of SEQ ID NO: 118 and the reverse primer of SEQ ID NO: 119; and / or the forward primer of SEQ ID NO: 120 and the reverse primer of SEQ ID NO: 121; and each fluorophore-quencher probe comprises a probe sequence selected from the group consisting of: SEQ ID NO: 39, SEQ ID NO: 40, SEQ ID NO: 96, SEQ ID NO: 97, SEQ ID NO: 98, SEQ ID NO: 99, SEQ ID NO: 99, SEQ ID NO: 100, SEQ ID NO: 101, SEQ ID NO: 102, SEQ ID NO: 103, SEQ ID NO: 104, SEQ ID NO: 105, SEQ ID NO: 106, SEQ ID NO: 107, SEQ ID NO: 118, SEQ ID NO: 119, and / or the forward primer of SEQ ID NO: 120 and the reverse primer of SEQ ID NO: 121; and each fluorophore-quencher probe comprises a probe sequence selected from the group consisting of: SEQ ID NO: 39, SEQ ID NO: 40, SEQ ID NO: 99, SEQ ID NO: 99, SEQ ID NO: 99, SEQ ID NO: ID NO: 41, SEQ ID NO: 43, SEQ ID NO: 44, SEQ ID NO: 45, SEQ ID NO: 48, SEQ ID NO: 52, SEQ ID NO: 55, SEQ ID NO: 122, SEQ ID NO: 123, SEQ ID NO: 124, SEQ ID NO: 125, SEQ ID NO: 126, SEQ ID NO: 127, SEQ ID NO: 128, SEQ ID NO: 134 and SEQ ID NO:

135.

155. A computer program product for determining the risk of or diagnosing in an individual inflammatory bowel disease (IBD) and / or an IBD subtype, wherein the IBD subtype is ulcerative colitis (UC) or Crohn's disease (CD), wherein the computer program product comprises a computer-readable medium encoded with computer-executable code, wherein the computer-executable code is configured to perform the steps of the method as described in any one of claims 2-47.

156. A computer program product for determining, diagnosing, or recommending treatment for inflammatory bowel disease (IBD) and / or IBD subtypes, wherein the IBD subtype is ulcerative colitis (UC) or Crohn's disease (CD), wherein the computer program product comprises a computer-readable medium encoded with computer-executable code, wherein the computer-executable code is configured to perform the steps of the method as described in any one of claims 2-47, and optionally further configured to perform the following steps: If an individual is determined to have an increased risk of developing IBD, UC, or CD, or is currently diagnosed with IBD, UC, or CD, appropriate IBD treatment, UC treatment, or CD treatment is recommended to that individual, respectively.

157. A computer program product for determining the risk or diagnosing an IBD subtype in an individual, wherein the IBD subtype is ulcerative colitis (UC) or Crohn's disease (CD), wherein the computer program product comprises a computer-readable medium encoded with computer-executable code, wherein the computer-executable code is configured to perform the steps of the method as described in any one of claims 48-78.

158. A computer program product for determining, diagnosing, or recommending treatment for an individual for an IBD subtype, wherein the IBD subtype is ulcerative colitis (UC) or Crohn's disease (CD), wherein the computer program product comprises a computer-readable medium encoded with computer-executable code, wherein the computer-executable code is configured to perform the steps of the method as described in any one of claims 48-78, and optionally further configured to perform the following steps: If an individual is determined to have an increased risk of developing IBD, UC, or CD, or is currently diagnosed with IBD, UC, or CD, appropriate IBD treatment, UC treatment, or CD treatment is recommended to that individual, respectively.

159. The method of any one of claims 79-112, wherein the first machine learning model and / or the second machine learning model is a generalized linear model.

160. The method of any one of claims 113-134, wherein the first machine learning model and / or the second machine learning model is a generalized linear model.

161. A method for determining, diagnosing, preventing, or treating Crohn's disease (CD) in an individual, said method comprising the steps of: a) Determine the relative abundance of a set of bacterial species in a fecal sample from the individual, wherein the set of bacterial species comprises one or more bacterial species selected from the group consisting of: formic acid-producing Dorperella, picky Eubacterium, inulin-eating Rosette, enteric Rosette, ovale-forming rumenococci (ovale-forming Broutella), protruding Eubacterium, anaerobic bacillus hardruci, sugar-insoluble Lawsonia, vibrio species CAG: 241, vibrio species 57_20, Lactobacillus rogueside, Eubacterium species CAG: 274, ileum Butz, Bacteroides fragilis, Escherichia coli, gingival groove Eubacterium, Actinomyces species oral taxonomy group 181 and Actinomyces species S6-Spd3; b) The relative abundance of each bacterial species in the bacterial species set from the individual is compared with the relative abundance of each bacterial species in the bacterial species set in the reference dataset to generate a risk score; c) If the risk score is higher than the cutoff value, the individual is determined to have an increased risk of developing CD or to already have CD; as well as d) If it is determined that the individual has an increased risk of developing CD or already has CD, the individual may optionally be treated.

162. A method for identifying fecal samples with altered Crohn's disease microbiome, the method comprising the steps of: a) Determine the relative abundance of a set of bacterial species in the fecal sample, wherein the set of bacterial species comprises one or more bacterial species selected from the group consisting of: formic acid-producing Dorperella, picky Eubacterium, inulin-eating Rosette, enteric Rosette, ovale-forming rumenococci (ovale-forming Broutella), protruding Eubacterium, anaerobic bacillus hardruci, unsaccharitroph Lawsonia, vibrio species CAG: 241, vibrio species 57_20, Lactobacillus rhamnosus, Eubacterium species CAG: 274, ileum Rombutz, Bacteroides fragilis, Escherichia coli, gingival groove Eubacterium, Actinomyces species oral taxonomy group 181 and Actinomyces species S6-Spd3; b) The relative abundance of each bacterial species in the bacterial species set from the individual is compared with the relative abundance of each bacterial species in the bacterial species set in the reference dataset to generate a risk score; as well as c) If the risk score is higher than the cutoff value, the fecal sample is determined to have an altered CD microbiome.

163. A method for increasing amino acid biosynthesis in cells, wherein the amino acid is L-arginine, L-ornithine, or L-valine, the method comprising providing the cells with an effective amount of one or more bacterial species selected from the group consisting of: ovomycosis (Broutella ovalis), inulin-eating Rosette, formic acid-producing Dorperella, Eubacterium species CAG: 274, and Enterobacter Rosette.

164. A method for increasing carbohydrate degradation in cells, the method comprising providing the cells with an effective amount of one or more bacterial species selected from the group consisting of: ovomycosis (ovomycosis broutitidis), inulin-eating Rosette, formic acid-producing Dorperella, Eubacterium species CAG: 274, and Enterobacter Rosette.

165. A method for increasing the biosynthesis of cofactors, carriers, and vitamins in cells, wherein the vitamin is thiamine pyrophosphate, the method comprising providing the cells with an effective amount of one or more bacterial species selected from the group consisting of: ovomycosis (Broutella ovalis), inulin-eating Rosette, formic acid-producing Dorperella, Eubacterium species CAG: 274, and Enterobacter Rosette.

166. A method for increasing amino acid biosynthesis in cells, wherein the amino acid is L-tryptophan, the method comprising providing the cells with an effective amount of ovoid rumenococci (ovoid broutella), inulin-eating Rosette, formic acid-producing Dorperella, Eubacterium species CAG: 274, or Enterobacter Rosette.

167. A composition comprising Lawsonia non-sugar-degrading bacteria and Eubacterium species CAG:

274.

168. A composition comprising ovomycosis (ovomycosis Broutella), Lawsonia non-sugar-degrading bacteria and inulin-eating Rossella.

169. A kit for determining the risk of or diagnosing Crohn's disease (CD) in an individual, the kit comprising reagents for detecting a set of bacterial species selected from the group consisting of: *Dolphobicella formatei*, *Eubacterium pickyii*, *Roseobacterium inulinum*, *Roseobacter rumenella*, *Eubacterium ovale*, *Eubacterium truncatella*, *Anaerobicus hardruns*, *Lawsonia suis*, *Oscillatoria* species CAG: 241, *Oscillatoria* species 57-20, *Lactobacillus rhamnosus*, *Eubacterium* species CAG: 274, *Romebuts ileum*, *Bacteroides fragilis*, *Escherichia coli*, *Eubacterium gingivalis*, *Actinomyces* species oral taxa 181, and *Actinomyces* species S6-Spd3.

170. A computer program product for determining the risk of or diagnosing Crohn's disease (CD) in an individual, wherein the computer program product comprises a computer-readable medium encoded with computer-executable code, wherein the computer-executable code is configured to perform the following steps: a) Determine the relative abundance of a set of bacterial species in a fecal sample from the individual, wherein the set of bacterial species comprises one or more bacterial species selected from the group consisting of: formic acid-producing Dorperella, picky Eubacterium, inulin-eating Rosette, enteric Rosette, ovale-forming rumenococci (ovale-forming Broutella), protruding Eubacterium, anaerobic bacillus hardruci, sugar-insoluble Lawsonia, vibrio species CAG: 241, vibrio species 57_20, Lactobacillus rogueside, Eubacterium species CAG: 274, ileum Butz, Bacteroides fragilis, Escherichia coli, gingival groove Eubacterium, Actinomyces species oral taxonomy group 181 and Actinomyces species S6-Spd3; b) Compare the relative abundance of each bacterial species in the bacterial species set from the individual with the relative abundance of each bacterial species in the bacterial species set in the reference dataset to generate a risk score; and c) If the risk score is higher than the cutoff value, the individual is determined to have an increased risk of developing CD or is already suffering from CD.

171. A method for determining, diagnosing, preventing, or treating Crohn's disease (CD) in an individual, said method comprising the steps of: a) Determine the relative abundance of a set of bacterial species in a fecal sample from said individual, said set of bacterial species comprising one or more bacterial species selected from the group consisting of: *Gimidae*, *Eubacterium hominis*, *Broutella ovalis*, *Roseobacterium inulinum*, *Clostridium sacchariformis*, *Eubacterium rectum*, *Roseobacterium falciparum*, *Bacillus sacchariformis*, *Collinsella aerogenes*, *Flavobacterium previae*, *Anaerobic bacillus hardrovirus*, *Trichophyton molasses*, *Ruminococcus tortifolius*, *Clostridium truncatula*, *Bacteroides falciparum*, *Ruminococcus broutii*, *Roseobacterium enterica*, *Adlerococcus faecium*, *Alternaria putrefactive*, and *Eubacterium* species CAG.

38. *Rhodotorula hygroscopica*, *Butyrate-producing Agastobacterium*, *Formicate-producing Dorl*, *Bacteroides faecium*, *Vibrio vulnificus*, *Oscillatoria* species 57_20, *Cercozoella fecalith*, *Bacteroides vulnificus*, *Dolberella longiformis*, *Alternaria sarcoptera*, *Lawsonia sucrase*, Firmicutes bacteria CAG: 83, *Oscillatoria* species CAG: 241, *Eubacterium filariae*, *Alternaria finniformis*, *Bacteroides faecium*, *Ackermania myxophilus*, *Associated cocci*, *Vibrio virulentis*, *Alternaria inconspicuous*, *Bacteroides monomorpha*, *Bacteroides faecium*, *Eubacterium* species CAG: 274, *Desulfovibrio suis*, *Bacteroides masei*, *Clostridium* species CAG:

299. *Eubacterium tumefaciens*, *Romebutzella ilei*, *Eubacterium tumefaciens*, *Anaerobic Bacillus hardruci*, *Lactobacillus rhamnosus*, *Ruminococcus ovalis*, *Twinococcus hemolyticus*, *Streptococcus pyogenes*, *Actinomyces cariogenicus*, *Actinomyces* species S6-Spd3, *Actinomyces griseus*, *Lactobacillus mucosae*, *Streptococcus parahaemolyticus*, *Romenothiazus*, *Actinomyces* species oral taxonomy group 181, *Veillonella septum*, *Escherichia coli*, *Bacteroides fragilis*, and *Eubacterium gingivalis*. b) The relative abundance of each bacterial species in the bacterial species set from the individual is compared with the relative abundance of each bacterial species in the bacterial species set in the reference dataset to generate a risk score; c) If the risk score is higher than the cutoff value, the individual is determined to have an increased risk of developing CD or to already have CD; as well as d) If it is determined that the individual has an increased risk of developing CD or already has CD, the individual may optionally be treated.

172. A method for monitoring Crohn's disease (CD) activity in an individual, the method comprising the steps of: a) Determine the relative abundance of a set of bacterial species in a fecal sample from the individual, wherein the set of bacterial species comprises one or more bacterial species selected from the group consisting of: formic acid-producing Dorperella, ovale-forming rumenococci (ovale-forming Broutella), sugar-insoluble Lawsonia, and Escherichia coli. b) The relative abundance of each bacterial species in the bacterial species set from the individual is compared with the relative abundance of each bacterial species in the bacterial species set in the reference dataset to generate a risk score; c) If the risk score is higher than the cutoff value, the individual is determined to have increased disease activity.

173. A method for determining, diagnosing, preventing, or treating ulcerative colitis (UC) in an individual, the method comprising the following steps: a) Determine the relative abundance of a set of bacterial species in a fecal sample from the individual, wherein the set of bacterial species comprises one or more bacterial species selected from the group consisting of: *Collagenia fecalis*, *Bacillus crypticarum*, *Colinus fecalis*, *Oscillatoria* species CAG: 241, *Lawsonia crypticarum*, *Toxococcus butyricum*, *Clostridium* species CAG: 58, *Eubacterium* species CAG: 274, *Bacillus warwickii*, *Clostridium truncatum*, *Clostridium sacchariformis*, *Gemmelia formica*, *Ostomyces viscerans*, *Ruminococcus tortifolia*, *Actinomyces* species oral taxonomy group 181, *Clostridium spiralis*, *Germococcus mesenteriae*, and *Broutella henselae*. b) The relative abundance of each bacterial species in the bacterial species set from the individual is compared with the relative abundance of each bacterial species in the bacterial species set in the reference dataset to generate a risk score; c) If the risk score is higher than the cutoff value, the individual is determined to have an increased risk of developing UC or already has UC; as well as d) If it is determined that the individual has an increased risk of developing UC or already has UC, the individual may optionally be treated.

174. A method for identifying a fecal sample with an altered ulcerative colitis microbiome, the method comprising the steps of: a) Determine the relative abundance of a set of bacterial species in the fecal sample, wherein the set of bacterial species comprises one or more bacterial species selected from the group consisting of: *Collagenia fecalis*, *Bacillus crypticarum*, *Colinus fecalis*, *Oscillatoria* species CAG: 241, *Lawsonia crypticarum*, *Toxococcus butyricum*, *Clostridium* species CAG: 58, *Eubacterium* species CAG: 274, *Bacillus warwickii*, *Clostridium truncatum*, *Clostridium sacchariformis*, *Gemmelia formica*, *Ostomyces viscerans*, *Ruminococcus tortifolia*, *Actinomyces* species oral taxonomy group 181, *Clostridium spiralis*, *Germococcus mesenteriae*, and *Broutella henselae*. b) The relative abundance of each bacterial species in the bacterial species set from the individual is compared with the relative abundance of each bacterial species in the bacterial species set in the reference dataset to generate a risk score; as well as c) If the risk score is higher than the cutoff value, the fecal sample is determined to have an altered UC microbiome.

175. A method for increasing nucleoside and nucleotide biosynthesis in cells, the method comprising providing the cells with an effective amount of one or more bacterial species selected from the group consisting of: Clostridium truncatum, Clostridium sacchariformis, Gemmaella formica, and Ruminococcus tortifolia.

176. A composition comprising Clostridium sacchariformis, Clostridium truncatum, and Formic acid-producing Gastroenterobacterium.

177. A kit for determining the risk of or diagnosing ulcerative colitis (UC) in an individual, said kit comprising reagents for detecting a set of bacterial species selected from the group consisting of: *Collagenia fecalis*, *Bacillus crypticarum*, *Colinus feces*, *Oscillatoria* species CAG: 241, *Lawsonia crypticarum*, *Toxococcus butyricum*, *Clostridium* species CAG: 58, *Eubacterium* species CAG: 274, *Bacillus warfarinii*, *Clostridium truncatum*, *Clostridium sacchariformis*, *Gemmelella formica*, *Ostomyces viscerans*, *Ruminococcus tortifolia*, *Actinomyces* species oral taxonomy group 181, *Clostridium spiralis*, *Germococcus mesenteriae*, and *Broutella henselae*.

178. A computer program product for determining the risk of or diagnosing ulcerative colitis (UC) in an individual, wherein the computer program product comprises a computer-readable medium encoded with computer-executable code, wherein the computer-executable code is configured to perform the following steps: a) Determine the relative abundance of a set of bacterial species in a fecal sample from the individual, wherein the set of bacterial species comprises one or more bacterial species selected from the group consisting of: *Collagenia fecalis*, *Bacillus crypticarum*, *Colinus fecalis*, *Oscillatoria* species CAG: 241, *Lawsonia crypticarum*, *Toxococcus butyricum*, *Clostridium* species CAG: 58, *Eubacterium* species CAG: 274, *Bacillus warwickii*, *Clostridium truncatum*, *Clostridium sacchariformis*, *Gemmelia formica*, *Ostomyces viscerans*, *Ruminococcus tortifolia*, *Actinomyces* species oral taxonomy group 181, *Clostridium spiralis*, *Germococcus mesenteriae*, and *Broutella henselae*. b) Compare the relative abundance of each bacterial species in the bacterial species set from the individual with the relative abundance of each bacterial species in the bacterial species set in the reference dataset to generate a risk score; and c) If the risk score is higher than the cutoff value, the individual is determined to have an increased risk of UC or is currently suffering from UC.

179. A method for determining, diagnosing, preventing, or treating ulcerative colitis (UC) in an individual, the method comprising the following steps: a) Determine the relative abundance of a set of bacterial species in a fecal sample from said individual, wherein said set of bacterial species comprises one or more bacterial species selected from the group consisting of: *Colinus oxyphylla*, *Clostridium truncatula*, *Ruminococcus tortifolia*, *Bacillus sacchariformis*, *Gemacobacterium formic acid*, *Clostridium sacchariformis*, *Alternaria putrefactiveis*, *Dolphinium longiformis*, *Ruminococcus brevicornu*, *Ostomyces viscerans*, *Trichophyton floccosum*, *Concomitant cocci*, *Adelaide cloacae*, *Roseidon*, *Brutella ovalis*, *Eubacterium rectum*, *Dolphinium formicate*, *Alternaria sarcopenia*, *Bacteroides fecalith*, *Bacteroides parafecalith*, and *Oscillatoria* species (CAG).

241. *Ackermania pseudomallei*, *Alternaria fibrinolytica*, *Eubacterium hopterii*, *Oscillatoria* species 57_20, *Rossiella humanis*, *Koala faecalis*, *Collinus faecalis*, *Bifidobacterium adolescentis*, *Rossiella enterica*, *Toxic Butyrosporum*, *Alternaria unobservatum*, *Eubacterium* species CAG: 38, *Anaerobic Bacillus hardrovirus*, *Bacteroides faecalis*, *Clostridium* species CAG: 242, *Lawsonia solani*, *Eubacterium* species CAG: 274, *Cercozoella fecalithii*, *Bacteroides masei*, *Rossiella faecalis*, *Clostridium* species CAG: 58, *Russobacterium lactis*, *Bacteroides vulgaris*, *Biliobacterium warfarinii*, Firmicutes bacteria CAG:

145. *Bacillus cereus*, *Enterobacter cecum*, *Twincocephalus measlesus*, *Veillonella infantis*, *Haemophilus* species HMSC71H05, *Actinomyces* species oral taxonomy group 181, *Blautella aerogenes*, *Lactobacillus mucosa*, *Enterococcus avium*, *Veillonella atypical*, *Blautella gingivalis*, *Blautella henseniella*, *Rhodopseudomonas myxobolus*, *Clostridium spiralis*, *Tizel*, *Veillonella septum*, and *Bacteroides fragilis*. b) The relative abundance of each bacterial species in the bacterial species set from the individual is compared with the relative abundance of each bacterial species in the bacterial species set in the reference dataset to generate a risk score; c) If the risk score is higher than the cutoff value, the individual is determined to have an increased risk of developing UC or already has UC; as well as d) If it is determined that the individual has an increased risk of developing UC or already has UC, the individual may optionally be treated.

180. A method for monitoring ulcerative colitis (UC) disease activity in an individual, the method comprising the steps of: a) Determine the relative abundance of a set of bacterial species in a fecal sample from the individual, wherein the set of bacterial species comprises one or more bacterial species selected from the group consisting of: Clostridium truncatum, Clostridium sacchariformis, Clostridium viscerum, Formicoblastic bacteria, Actinomyces species oral taxonomy 181, and Clostridium spiralis. b) The relative abundance of each bacterial species in the bacterial species set from the individual is compared with the relative abundance of each bacterial species in the bacterial species set in the reference dataset to generate a risk score; as well as c) If the risk score is higher than the cutoff value, the individual is determined to have increased disease activity.

181. A computer program product for determining, diagnosing, or recommending treatment for Crohn's disease (CD) in an individual, wherein the computer program product comprises a computer-readable medium encoded with computer-executable code, wherein the computer-executable code is configured to perform the following steps: a) Determine the relative abundance of a set of bacterial species in a fecal sample from the individual, wherein the set of bacterial species comprises one or more bacterial species selected from the group consisting of: formic acid-producing Dorperella, picky Eubacterium, inulin-eating Rosette, enteric Rosette, ovale-forming rumenococci (ovale-forming Broutella), protruding Eubacterium, anaerobic bacillus hardruci, sugar-insoluble Lawsonia, vibrio species CAG: 241, vibrio species 57_20, Lactobacillus rogueside, Eubacterium species CAG: 274, ileum Butz, Bacteroides fragilis, Escherichia coli, gingival groove Eubacterium, Actinomyces species oral taxonomy group 181 and Actinomyces species S6-Spd3; b) The relative abundance of each bacterial species in the bacterial species set from the individual is compared with the relative abundance of each bacterial species in the bacterial species set in the reference dataset to generate a risk score; c) If the risk score is higher than the cutoff value, the individual is determined to have an increased risk of developing CD or already has CD; and d) If it is determined that the individual has an increased risk of developing CD or already has CD, then the individual may optionally be advised of appropriate CD treatment.

182. A computer program product for determining the risk of or diagnosing ulcerative colitis (UC) in an individual, wherein the computer program product comprises a computer-readable medium encoded with computer-executable code, wherein the computer-executable code is configured to perform the following steps: a) Determine the relative abundance of a set of bacterial species in a fecal sample from the individual, wherein the set of bacterial species comprises one or more bacterial species selected from the group consisting of: *Collagenia fecalis*, *Bacillus crypticarum*, *Colinus fecalis*, *Oscillatoria* species CAG: 241, *Lawsonia crypticarum*, *Toxococcus butyricum*, *Clostridium* species CAG: 58, *Eubacterium* species CAG: 274, *Bacillus warwickii*, *Clostridium truncatum*, *Clostridium sacchariformis*, *Gemmelia formica*, *Ostomyces viscerans*, *Ruminococcus tortifolia*, *Actinomyces* species oral taxonomy group 181, *Clostridium spiralis*, *Germococcus mesenteriae*, and *Broutella henselae*. b) The relative abundance of each bacterial species in the bacterial species set from the individual is compared with the relative abundance of each bacterial species in the bacterial species set in the reference dataset to generate a risk score; c) If the risk score is higher than the cutoff value, the individual is determined to have an increased risk of developing UC or currently has UC; and If it is determined that the individual has an increased risk of developing UC or already has UC, appropriate UC treatment may be recommended to the individual.

183. The method of any one of claims 2-47, wherein the first set of bacterial species and the second set of bacterial species each independently comprise two or more bacterial species selected from the group consisting of: Actinomycetes species oral taxa 181, Bacteroides fragilis, Escherichia coli, Lawsonia sucrase, Eubacterium species CAG: 274, Rosette of inulin, Rosette of the gut, Oval rumenococcus (Oval Brutella), Formic acid-producing Dorl, Bileophilus warwickii, Clostridium truncatella, Clostridium sucrase, Formic acid-producing Gimeria, Visceral bacterium, Ruminococcus tortiformis, Clostridium spiralis, Gemmaella muscarinae, and Broutella hensenae.

184. The method of any one of claims 48-78, wherein the first set of bacterial species and the second set of bacterial species each independently comprise two or more bacterial species selected from the group consisting of: Actinomycetes species oral taxa 181, Bacteroides fragilis, Escherichia coli, Lawsonia sucrase, Eubacterium species CAG: 274, Rosette of inulin, Rosette of the gut, Oval rumenococcus (Oval Brutella), Formic acid-producing Dorl, Bileophilus warwickii, Clostridium truncatella, Clostridium sucrase, Formic acid-producing Gemmaella, Visceral Odorobacter, Ruminococcus tortiformis, Clostridium spiralis, Gemmaella muscarinae, and Broutella hensenae.

185. The method of any one of claims 79-112, wherein the first set of bacterial species and the second set of bacterial species each independently comprise two or more bacterial species selected from the group consisting of: Actinomycetes species oral taxa 181, Bacteroides fragilis, Escherichia coli, Lawsonia sucrase, Eubacterium species CAG: 274, Rosette of inulin, Rosette of the gut, Oval rumenococcus (Oval Brutella), Formic acid-producing Dorl, Bileophilus warwickii, Clostridium truncatella, Clostridium sucrase, Formic acid-producing Gemmaella, Visceral Odorobacter, Ruminococcus tortiformis, Clostridium spiralis, Gemmaella muscarinae, and Broutella hensenae.

186. The method of any one of claims 113-160, wherein the first set of bacterial species and the second set of bacterial species each independently comprise two or more bacterial species selected from the group consisting of: Actinomycetes species oral taxa 181, Bacteroides fragilis, Escherichia coli, Lawsonia sucrase, Eubacterium species CAG: 274, Rosette of inulin, Rosette of the gut, Oval rumenococcus (Oval Brutella), Formic acid-producing Dorl, Bileophilus warwickii, Clostridium truncatella, Clostridium sucrase, Formic acid-producing Gemmaella, Visceral Odorobacter, Ruminococcus tortiformis, Clostridium spiralis, Gemmaella muscarinae, and Broutella hensenii.

Images