Method for diagnosing migraine using oral microbiome

Abstract

The present disclosure relates to a method for providing information for diagnosing migraine and, more specifically, to a method for accurately and quickly diagnosing migraine without pain or discomfort to a patient, the method comprising a step of measuring the relative abundance of at least one bacterium selected from the group consisting of Gemella haemolysans, Centipeda periodontii, Cardiobacterium valvarum, Catonella morbi, and Selenomonas noxia in a biological sample.

Description

Diagnosing migraine using the oral microbiome

[0001] Cross-reference to related applications

[0002] This disclosure claims the benefit of priority to Republic of Korea Patent Application No. 10-2024-0097876, filed July 24, 2024, Republic of Korea Patent Application No. 10-2025-0060463, provisional filed May 9, 2025, and patent application No. 10-2025-0099887, filed July 23, 2025, each of which is incorporated by reference herein in its entirety.

[0003] The present disclosure relates to a method for diagnosing migraine using oral microbiome.

[0004] Migraine is a primary headache disorder characterized by recurrent headache attacks due to abnormalities in the nerves and blood vessels.

[0005] Migraines are characterized by severe headaches, often accompanied by symptoms such as nausea, vomiting, and photophobia, and are often severe. Therefore, those suffering from migraines typically experience significant disability, including a reduced quality of life, absences from school or work, and decreased productivity.

[0006] Furthermore, migraine is known to have a global prevalence of approximately 15%. According to the 2015 Global Burden of Disease (GBD) study conducted by the World Health Organization, migraine accounted for 5.61% of all disease-related years lived with disability (YLD), ranking it as the sixth leading cause of disability. In particular, it was ranked as the third leading cause of disability among those aged 15-49, a highly socially active age group. Thus, the burden of migraine on individuals and society is enormous.

[0007] Migraine is a prevalent and serious neurological disorder that affects more than 1 billion people, significantly impacting individuals' quality of life and placing a significant burden on healthcare systems. 1, 2 Although some key mediators, such as calcitonin gene-related peptide (CGRP), have been shown to play a role in migraine, the full pathophysiology remains unknown. 3, 4 The lack of migraine biomarkers often delays diagnosis, leads to suboptimal treatment, and increases the burden of migraine on patients.

[0008] Meanwhile, the diagnosis of migraine is based on the patient's medical history and symptoms, and the criteria established by the International Headache Society are generally used. The lack of objective diagnostic testing methods for migraine limits our understanding of the pathophysiology of migraine and the development of migraine treatments.

[0009] Accordingly, the inventors of the present invention conducted extensive research to develop a method for diagnosing migraines and determining their types, and as a result, they confirmed the correlation between the oral microbiome and migraines, thereby completing the present invention.

[0010] Emerging evidence suggests that oral health conditions can exacerbate migraines, and that saliva is a potential source of migraine biomarkers. The triple interaction of the oral-gut-brain axis is known to be associated with various neurological disorders, but little research has been conducted on migraine.

[0011] A growing body of research suggests that saliva contains a number of potential biomarkers for migraine, which could provide valuable insights into the physiological changes observed in individuals with migraine. 5-11 In particular, recent studies on salivary CGRP have highlighted the potential of saliva as a vehicle for investigating biomarkers relevant to migraine diagnosis and treatment.10, 11

[0012] In contrast, emerging evidence suggests that the gut-brain axis, and by extension the oral-gut-brain axis, may influence the pathogenesis of several neurological disorders. 12-15 The gut-brain axis is a two-way communication between the gastrointestinal system and the central nervous system (CNS). 12 It is greatly influenced by the gut microbiome, neurotransmitters, and metabolites such as short-chain fatty acids (SCFAs). 16

[0013] This concept extends to the oral-gut-brain axis, including the oral microbiome, which is the second largest microbiome after the gut and is mechanically and chemically linked to the gut. 17 Moreover, the oral microbiome and its byproducts directly affect the brain via the trigeminal, olfactory, facial nerves, and bloodstream. 13, 18 Alternatively, they may indirectly affect the brain by transporting microbes from saliva to the gut, causing systemic inflammation, or by disrupting the gut microbiota composition. 17, 19

[0014] Disruptions in the balance of the oral and gut microbiome may be linked to the pathogenesis of migraine. Some previous studies have shown changes in the oral or gut microbiome in migraine patients. 20-23 However, the heterogeneity of research results makes it difficult to identify specific factors contributing to migraine. Therefore, there is a need to simultaneously investigate the oral and gut microbiomes and analyze their complex interrelationships.

[0015] We simultaneously investigated the oral and gut microbiota imbalances in patients with episodic migraine (EM) and chronic migraine (CM) compared with healthy controls (HC), and further identified potential taxonomic and functional characteristics of migraine using the oral microbiome.

[0016] Based on the results of simultaneous investigation of the oral and intestinal microbiomes, the present inventors further expanded and developed the present invention by identifying key oral microbiomes (signature microorganisms) that may influence migraine.

[0017] [Prior Art Literature]

[0018] Headache Classification Committee of the International Headache Society (IHS) The International Classification of Headache Disorders, 3rd edition. Cephalalgia. 2018; 38: 1-211.

[0019] The present disclosure aims to provide a method for diagnosing migraine using oral microbiome.

[0020] The present disclosure aims to provide a method for determining migraine severity using the oral microbiome.

[0021] The present disclosure aims to provide a method for diagnosing episodic migraine (EM) using the oral microbiome.

[0022] The present disclosure aims to provide a method for diagnosing chronic migraine (CM) using the oral microbiome.

[0023] The present disclosure aims to provide a method and device for diagnosing and / or determining the type of migraine by utilizing a machine learning model trained on oral microbiome features.

[0024] The present disclosure provides the following embodiments as means or solutions for solving the above-mentioned technical problems:

[0025] Embodiment 1. A method for providing information for diagnosing migraine, comprising the step of measuring the relative abundance of at least one microorganism selected from the group consisting of the genera Megasphaera, Catonella, Haemophilus and Selenomonas in a biological sample obtained from an individual.

[0026] Embodiment 2. A method for providing information for diagnosing migraine, further comprising the step of measuring the relative abundance of at least one bacteria selected from the group consisting of the genus Gemella, the genus Granulicatella and the genus Streptococcus in the biological sample, in accordance with Embodiment 1.

[0027] Embodiment 3. A method for providing information for diagnosing migraine, wherein in Embodiment 1, a diagnosis of migraine is made when the relative abundance of at least one microorganism selected from the group consisting of the genus Megasphaera, the genus Catonella, the genus Haemophilus, and the genus Selenomonas is decreased compared to a normal control group.

[0028] Embodiment 4. A method for providing information for diagnosing migraine, wherein in Embodiment 2, a diagnosis of migraine is made when the relative abundance of at least one microorganism selected from the group consisting of the genus Gemela, the genus Granulocatella, and the genus Streptococcus is increased compared to a normal control group.

[0029] Embodiment 5. A method for providing information for diagnosing migraine, further comprising the step of measuring the relative abundance of at least one microorganism selected from the group consisting of the genera Veillonella, Centipeda, Kingella, Campylobacter, Prevotella, Cardiobacterium, and Alloprevotella in the biological sample.

[0030] Embodiment 6. A method for providing information for diagnosing migraine, wherein in Embodiment 5, a diagnosis of migraine is made when the relative abundance of at least one microorganism selected from the group consisting of the genus Veillonella, the genus Centipeda, the genus Kingella, the genus Campylobacter, the genus Prevotella, the genus Cardiobacterium, and the genus Alloprevotella is decreased compared to a normal control group.

[0031] Embodiment 7. A method for providing information for diagnosing migraine, wherein the biological sample is saliva in Embodiment 1.

[0032] Embodiment 8. A method for providing information for diagnosing migraine, wherein, in Embodiment 5, episodic migraine is diagnosed when the relative abundance of microorganisms of the genus Prevotella is reduced compared to a normal control group, and the microorganisms of the genus Prevotella include at least one selected from the group consisting of Prevotella oris and Prevotella salivae.

[0033] Embodiment 9. A method for providing information for diagnosing migraine, further comprising the step of measuring the relative abundance of bacteria of the genus Porphyromonas in the biological sample, in Embodiment 1.

[0034] Embodiment 10. A method for providing information for diagnosing migraine, wherein, in Embodiment 9, episodic migraine is diagnosed when the relative abundance of microorganisms of the genus Porphyromonas is increased compared to a normal control group.

[0035] Embodiment 11. A method for providing information for diagnosing migraine, wherein in Embodiment 5, chronic migraine is diagnosed when the relative abundance of microorganisms of the genus Prevotella is reduced compared to a normal control group, and the microorganisms of the genus Prevotella include at least one selected from the group consisting of Prevotella saccharolytica, Prevotella pallens, and Prevotella nanceiensis.

[0036] Embodiment 12. A method for providing information for diagnosing migraine, wherein in Embodiment 5, if the relative abundance of microorganisms of the genus Cardiobacterium is reduced compared to a normal control group and the microorganisms of the genus Cardiobacterium are Cardiobacterium hominis, chronic migraine is diagnosed.

[0037] Embodiment 13. A method for providing information for diagnosing migraine, further comprising the step of measuring the relative abundance of at least one microorganism selected from the group consisting of the genus Rothia, the genus Mogibacterium and the genus Schaalia in the biological sample.

[0038] Embodiment 14. A method for providing information for diagnosing migraine, wherein, in Embodiment 13, chronic migraine is diagnosed when the relative abundance of at least one microorganism selected from the group consisting of the genus Rosia, the genus Mosquitobacterium, and the genus Shaalia is increased compared to a normal control group.

[0039] Embodiment 15. A method for providing information for diagnosing migraine, further comprising the step of measuring the relative abundance of at least one microorganism selected from the group consisting of the genera Fusobacterium, Haemophilus and Neisseria in the biological sample.

[0040] Embodiment 16. A method for providing information for diagnosing migraine, wherein, in Embodiment 15, a diagnosis of chronic migraine is made when the relative abundance of at least one microorganism selected from the group consisting of the genus Fusobacterium, the genus Haemophilus, and the genus Neisseria is decreased.

[0041] Embodiment 17. A method for diagnosing and / or determining migraine using a machine learning model performed on a device,

[0042] A step of extracting multiple microbial data from a biological sample collected from an object;

[0043] A step of selecting microbial-related features to be used in a machine learning model from among a plurality of extracted microbial data based on a preset feature selection algorithm;

[0044] A step of training a machine learning model using selected microbial-related features; and

[0045] A step of diagnosing and / or classifying migraine by inputting microbial data collected from the subject to be examined into the learned machine learning model,

[0046] The above microbial related features are:

[0047] The total content and relative abundance of at least one microorganism selected from a first group of microorganisms consisting of microorganisms of the genus Gemella, microorganisms of the genus Streptococcus, microorganisms of the genus Granulicatella, microorganisms of the genus Rothia, microorganisms of the genus Mogibacterium, and microorganisms of the genus Schaalia; and

[0048] A method comprising the total content and relative abundance of at least one microorganism selected from a second group of microorganisms consisting of microorganisms of the genus Alloprevotella, microorganisms of the genus Veillonella, microorganisms of the genus Haemophilus, microorganisms of the genus Selenomonas, microorganisms of the genus Campylobacter, microorganisms of the genus Cardiobacterium, microorganisms of the genus Megasphaera, and microorganisms of the genus Kingella.

[0049] Embodiment 18. In Embodiment 17, the machine learning model is selected from among a linear regression analysis (LRA) model, a random forest model, a generalized linear (GLMNET) model, a gradient boosting model, and an extreme gradient boost (XGB) model.

[0050] Embodiment 19. In Embodiment 17, a method for diagnosing migraine, wherein, in a plurality of microbial data extracted from microorganisms collected from the subject to be examined, the relative abundance of at least one microorganism selected from the first microbial group increases and / or the relative abundance of at least one microorganism selected from the second microbial group decreases.

[0051] The method for providing information for diagnosing migraine according to the present disclosure enables accurate diagnosis / identification of migraine.

[0052] The method for providing information for diagnosing migraine according to the present disclosure enables rapid diagnosis / identification of migraine.

[0053] The method for providing information for diagnosing migraine according to the present disclosure is a non-invasive method that enables diagnosis / discernment of migraine without causing pain or rejection in the subject (patient).

[0054] The method for providing information for diagnosing migraine according to the present disclosure can be used to determine the severity of migraine.

[0055] The method for providing information for diagnosing migraine according to the present disclosure enables specific diagnosis / discrimination by distinguishing between episodic migraine and chronic migraine.

[0056] Oral microorganisms and combinations thereof according to the present disclosure are useful as biomarkers for diagnosing and / or identifying migraine and can be utilized as therapeutic targets.

[0057] Figure 1 shows the types of microorganisms that showed differences in the oral microbiome of migraine patients compared to the normal control (HC) group.

[0058] Figures 2a-2b show the results of differential abundance tests of the oral microbiomes of the healthy control (HC) group and migraine patients. Figure 2a shows the species that showed significant differences when comparing the oral microbiomes of the chronic migraine (CM) group and the healthy control (HC) group, and Figure 2b shows the species that showed significant differences when comparing the oral microbiomes of the episodic migraine (EM) group and the healthy control (HC) group. Differential abundance tests were performed using the SIAMCAT R package, and the analysis included preprocessing options with a prevalence filtering cutoff of 0.1 and multiple testing correction using the Benjamini-Hochberg procedure with a significance cutoff of 0.05.

[0059] Figures 3a-3c show the results of verifying the performance of a method for providing information for diagnosing migraine according to one embodiment of the present disclosure. Figure 3a shows the classification results of migraine (chronic migraine (CM) and episodic migraine (EM)) and healthy control (HC) groups, Figure 3b shows the classification results of CM and HC groups, and Figure 3c shows the classification results of episodic migraine (EM) and healthy control (HC) groups. Figures 3a to 3c show the probability of a validation set obtained through 5-repeated 5-fold cross-validation of a logistic regression model according to machine learning-based classification of migraine conditions using oral microbial signatures, the receiver operating characteristic (ROC) curve of the entire data resampled from the response variable class, and the area under the curve (AUC).

[0060] Figures 4a-4b show the diversity of the oral and gut microbiomes in chronic migraine (CM), episodic migraine (EM), and healthy control (HC) groups. The Shannon diversity index and alpha diversity based on the number of taxa are shown for the oral group (a) and gut microbiome group (b). ****p<0.0001; *p<0.05.

[0061] Figure 5a shows differences in oral microbiome abundance at the genus level between the chronic migraine (CM) group and the healthy control (HC) group, and Figure 5b shows differences in oral microbiome abundance at the genus level between the episodic migraine (EM) group and the healthy control (HC) group. The differences in microbial genera, significance, fold change, prevalence change, and classification performance (AUC) of microbial markers at the genus level are presented using the SIAMCAT R package. Figure 5c shows a Venn diagram of oral microbial signatures that were increased or decreased in the chronic migraine (CM) and / or episodic migraine (EM) group compared to the healthy control (HC) group. Taxa enriched in migraine patients, represents taxa enriched in the control group.

[0062] Figures 6a-6b show differences in the abundance of functional metabolic pathway profiles in the oral microbiome. Figure 6a: Differences in abundance between chronic migraine (CM) and healthy control (HC) groups; Figure 6b: Differences in abundance between episodic migraine (EM) and healthy control (HC) groups. In the volcano plot, the x-axis represents the fold change (FC) between the two comparison groups, and the y-axis represents the p-value. Features with adjusted p < 0.05 are indicated by different symbols (empty / filled circles, empty / filled triangles, empty / filled squares, and asterisks) according to the pathway class, and significant features with FC ≥ 0.1 are described in text.

[0063] Figure 7a shows the relative proportions of ecological structural types between groups for the oral microbiome; Figure 7b shows a PCoA plot based on Bray-Curtis dissimilarity according to the ecological structure of the oral microbiome; Figure 7c shows the results of LEfSe analysis comparing genus profiles of the oral microbiome by ecological structure; Figure 7d shows the relative proportions of ecological structural types between groups for the gut microbiome; Figure 7e shows a PCoA plot based on Bray-Curtis dissimilarity according to the ecological structure of the gut microbiome; Figure 7f shows the results of LEfSe analysis comparing genus profiles of the gut microbiome by ecological structure.

[0064] Figure 8a shows the mean relative abundance (hatched) of oral microbial ASVs (amplicon sequence variants) found in both the oral cavity and the gut, and the mean relative abundance of ASVs detected only in the oral cavity; Figure 8b shows the mean relative abundance (hatched) of gut microbial ASVs found in both the oral cavity and the gut, and the mean relative abundance of ASVs detected only in the gut; Figure 8c shows the proportion of participants with shared ASVs between oral and gut paired samples in the chronic migraine (CM), episodic migraine (EM), and healthy control (HC) groups, respectively, for the eight most commonly co-occurring microbial genera.

[0065] Figures 9a-9c show the results of evaluating the performance of a machine learning model for migraine classification built using three feature sets (host, microbiome, host and microbiome features) based on a random forest modeling technique according to one embodiment of the present disclosure. Figure 9a: Migraine; Figure 9b: Chronic migraine (CM); Figure 9c: Episodic migraine (EM). In this machine learning model, eight factors extracted from lifestyle factors including age, sex, body mass index, and dietary habits were used as principal components as host features, and oral microbiome signatures were included as microbiome features. For each data set, 5-fold cross-validation was randomly repeated 10 times, and the performance of the classification model was evaluated using the area under the receiver operating characteristic (ROC) curve (AUC), and feature importance analysis for the model was performed using microbiome features. Figures 9d-9f are bar graphs visualizing the average feature importance values ​​corresponding to migraine (d), chronic migraine (CM) (e), and episodic migraine (EM) (f) for a classification model based solely on oral microbiome features.

[0066] Figure 10 shows the results of sample size calculation based on alpha diversity. The sample size was calculated using data obtained from our previous study [Sci Rep. 2023 Jan 12;13(1):626]. After calculating the mean and standard deviation of the Shannon index for each group, the difference values ​​were set to 0.3, 0.4, and 0.5, and the effect size according to the change in the difference value was calculated to estimate the number of samples accordingly. The effect sizes for the difference values ​​of 0.3, 0.4, and 0.5 were 0.68, 0.91, and 1.14, respectively. In order to confirm a statistically significant difference (α < 0.05, power > 80%) in alpha diversity between the three groups, it was estimated that a total of 70 samples (at least 24 per group) were required based on the effect size of 0.68.

[0067] Figure 11 shows the results of sample size calculations based on beta diversity. The sample size was calculated using data obtained from our previous study [Sci Rep. 2023 Jan 12;13(1):626]. The between-group effect size was estimated based on the observed distribution, and the total sample size required to achieve a target power of 0.8 was obtained through power curves according to significance levels (α = 0.001, 0.01, 0.05, and 0.1).

[0068] Figure 12 shows principal component analysis (PCA) of the lifestyle survey data. Figure 12a: Scree plot for selecting the number of PCs to be used in subsequent modeling. Figure 12b: Percentage of explained variance for each PC axis, with PCs 1 to 8 highlighted. Figures 12ca-12ch: PC loadings of the top 20 variables for each axis. Figures 12da-12dh: Comparison of PC axes by group based on Wilcoxon rank-sum tests. **P<0.01; *P<0.05; ns, not significant.

[0069] Figures 13a-13i show the association between monthly migraine frequency and oral microbiome signatures after adjusting for age, sex, BMI, lifestyle factors (PC 1-8), anxiety, and depression as covariates based on Poisson regression analysis.

[0070] Figure 14 shows the levels of acetic acid (a), propionic acid (b), and butyric acid (c) in the serum of chronic migraine (CM), episodic migraine (EM), and normal control (HC) groups. ****P< 0.0001; ***P< 0.001; ns, not significant.

[0071] Figures 15a-15c illustrate classification of migraine based on a machine learning model according to an embodiment of the present disclosure using all oral microbiome features after prevalence filtering of the entire oral microbiome using a cutoff of less than 10%: (a) Migraine, (b) Chronic Migraine (CM), and (c) Episodic Migraine (EM). Figures 15d-15f illustrate mean feature importance values ​​of the migraine classification model according to the present disclosure as bar graphs, with oral microbiome signatures highlighted as dots: (da-db) Migraine, (ea-eb) Chronic Migraine (CM), and (fa-fb) Episodic Migraine (EM).

[0072] Hereinafter, embodiments of the present disclosure will be described in detail with reference to the accompanying drawings so that those skilled in the art can easily implement the present disclosure. However, the present disclosure may be implemented in various different embodiments and is not limited to the embodiments described herein. Like reference numerals designate similar parts throughout the specification.

[0073] Additionally, when the terms “about,” “approximately,” or similar expressions such as “at least” are used in connection with a numerical value in this disclosure, it is intended that a theoretical, experimental, statistical, or empirical error of ±20%, ±10%, ±7%, ±5%, ±3%, ±2%, or ±1%, typically ±5%, be allowed based on that numerical value.

[0074] In one aspect, the present disclosure relates to a method of providing information for diagnosing migraine.

[0075] A method for providing information for diagnosing migraine according to the present disclosure comprises the step of measuring the relative abundance of at least one microorganism selected from the group consisting of the genera Megasphaera, Cartonella, Haemophilus and Selenomonas in a biological sample obtained from an individual.

[0076] In the present disclosure, an individual means any animal, such as livestock and mice, and may be a mammal, including humans, for example.

[0077] In one embodiment, the entity may be a human.

[0078] In the present disclosure, a biological sample may be any material derived from an individual, such as saliva, sweat, gastrointestinal fluid, tears, serum, urine, or feces.

[0079] In one embodiment, the biological sample may be saliva.

[0080] In the present disclosure, relative abundance means the relative species abundance in the microbiome, i.e., the ratio of the microbial composition of a specific species to the total number of microbial species.

[0081] In this disclosure, the microbiome refers to the total community of all microorganisms existing in human tissues and biological fluids, such as the gastrointestinal tract, skin, mammary glands, semen, uterus, follicles, lungs, saliva, oral mucosa, conjunctiva, and biliary tract, and the corresponding anatomical sites where they exist. Microorganisms include microorganisms, archaea, fungi, protozoa, and viruses.

[0082] In the present disclosure, the step of measuring the relative abundance of microorganisms is a step of detecting a microbiome in a biological sample and measuring the relative abundance of microorganisms constituting the microbiome.

[0083] In the present disclosure, the step of measuring the relative abundance of microorganisms may include the following steps:

[0084] (1) a step of amplifying a microbial gene from a biological sample; and

[0085] (2) Step of analyzing the base sequence of the amplified gene.

[0086] In the present disclosure, the step of measuring the relative abundance of microorganisms can be performed using primers, probes, antisense oligonucleotides, LNA (Locked Nucleic Acids), aptamers or antibodies that can specifically detect organic biomolecules such as proteins, nucleic acids, lipids, glycolipids, glycoproteins or sugars that are specifically present in microorganisms.

[0087] In one embodiment, the step of measuring the relative abundance of the microorganism can be performed by any one selected from the group consisting of polymerase chain reaction (PCR), real-time RT-PCR, RNase protection assay (RPA), northern blotting, and DNA chip using primers, probes, antisense oligonucleotides, or LNAs having a sequence complementary to 16s rRNA of the microorganism.

[0088] In this disclosure, migraine diagnosis means determining the likelihood of a subject developing migraine, whether a subject has developed migraine, or, if a subject has developed migraine, the severity of the condition.

[0089] In the present disclosure, a diagnosis of migraine can be made when the relative abundance of microorganisms in a biological sample is different compared to a control group, e.g., an individual who has never suffered from migraine or has suffered from migraine but has not experienced migraine symptoms for at least 3 months, preferably at least 6 months.

[0090] In the present disclosure, microorganisms whose relative abundance in biological samples collected from migraine patients is different compared to the control group include microorganisms of the genera Megasphaera, Catonella, Haemophilus, Selenomonas, Gemmella, Granulicatella, Streptococcus, Veillonella, Centipeda, Kingella, Campylobacter, Prevotella, Cardiobacterium, Alloprevotella, Porphyromonas, Rossia, Mosquitobacterium, Shaalia, Fusobacterium, Haemophilus, and Neisseria.

[0091] In one embodiment, a diagnosis of migraine can be made when the relative abundance of at least one microorganism selected from the group consisting of the genera Megasphaera, Catonella, Haemophilus, and Selenomonas in a biological sample obtained from an individual is decreased compared to a normal control group.

[0092] In one embodiment, a diagnosis of migraine can be made when the relative abundance of at least one microorganism selected from the group consisting of the genera Gemela, Granulicatella, and Streptococcus in a biological sample obtained from an individual is increased compared to a normal control group.

[0093] In one embodiment, a diagnosis of migraine can be made when the relative abundance of at least one microorganism selected from the group consisting of the genera Veillonella, Centipeda, Kingella, Campylobacter, Prevotella, Cardiobacterium, and Alloprevotella in a biological sample obtained from an individual is decreased compared to a normal control group.

[0094] In one embodiment, episodic migraine can be diagnosed when the relative abundance of Prevotella oris and / or Prevotella salibae is decreased in a biological sample obtained from an individual or migraine patient compared to a normal control group.

[0095] In one embodiment, episodic migraine can be diagnosed when the relative abundance of microorganisms of the genus Porphyromonas is increased in a biological sample obtained from an individual or a migraine patient compared to a normal control group.

[0096] In one embodiment, a diagnosis of chronic migraine can be made when at least one selected from the group consisting of Prevotella saccharolytica, Prevotella phalens and Prevotella nanseiensis is decreased in a biological sample obtained from an individual or a migraine patient compared to a normal control group.

[0097] In one embodiment, a diagnosis of chronic migraine can be made when the relative abundance of Cardiobacterium hominis is decreased in a biological sample obtained from an individual or a migraine patient compared to a normal control group.

[0098] In one embodiment, a diagnosis of chronic migraine can be made when the relative abundance of at least one microorganism selected from the group consisting of the genus Rossia, the genus Mosquitobacterium, and the genus Shaalia is increased in a biological sample obtained from an individual or a migraine patient compared to a normal control group.

[0099] In one embodiment, a diagnosis of chronic migraine can be made when the relative abundance of at least one microorganism selected from the group consisting of the genera Fusobacterium, Haemophilus, and Neisseria is decreased in a biological sample obtained from an individual or a migraine patient.

[0100] In another aspect, the present disclosure relates to a method of diagnosing and / or identifying migraine using a machine learning model performed on a device.

[0101] In one embodiment, a method for diagnosing and / or identifying migraine using a machine learning model according to the present disclosure comprises the following steps:

[0102] A step of extracting multiple microbial data from a biological sample collected from an object;

[0103] A step of selecting microbial-related features to be used in a machine learning model from among a plurality of extracted microbial data based on a preset feature selection algorithm;

[0104] A step of training a machine learning model using selected microbial-related features; and

[0105] A step of diagnosing and / or classifying migraine by inputting microbial data collected from the subject to be examined into the learned machine learning model.

[0106] In one embodiment, the device may include a personal computer such as a desktop or laptop, as well as a mobile terminal capable of wired or wireless communication. The mobile terminal is a wireless communication device that ensures portability and mobility, and may include various devices equipped with communication modules such as a smartphone, tablet PC, wearable device, and Bluetooth (BLE, Bluetooth Low Energy), NFC, RFID, ultrasonic, infrared, WiFi, LiFi, etc.

[0107] The device of the present disclosure may include a microbial data extraction unit, a microbial-related feature selection unit, a learning unit, and a diagnosis / discrimination unit.

[0108] The device of the present disclosure can diagnose or classify the presence and / or type of migraine based on variations (relative changes) in the microbiome, particularly the signature microbiome, in a sample collected from an individual.

[0109] The microbial data extraction unit can extract multiple microbial data sets from samples collected from an individual. Here, the multiple microbial data sets can be classified into training data (training set) and test data (test set) for learning, and the classification ratios can vary. When configured for k-fold cross-validation, the data is evenly divided into k sets. For each iteration, one set is used as the validation set (test set), and the remaining k-1 sets are used as the training set. Five-fold cross-validation consists of a 20% test set and an 80% training set.

[0110] The microbiome can be analyzed using genome-based analysis methods such as real-time PCR using microbial-specific primers suggested in the GULDA method or metagenome analysis using next-generation sequencing after extracting all genomes in the sample.

[0111] The microbial feature selection unit can select signature microbial variables from multiple microbial data sets (i.e., feature selection) to be used in a machine learning model based on a preset selection algorithm. Microbial variables may include microbial species, concentration, total content, and relative abundance.

[0112] Various variables (features, or variables, attributes) can be used to build a machine learning model, but if too many variables are used or inappropriate variables are used, the machine learning model may overfit or the prediction accuracy may decrease.

[0113] Therefore, for machine learning models to achieve high predictive accuracy, they need to use an appropriate combination of variables. By selecting the variables most closely correlated with the variable being predicted, the complexity of the machine learning model can be reduced while using as few variables as possible.

[0114] Variable selection algorithms that can be used include, for example, the Boruta algorithm and the Recursive Feature Elimination (RFE) algorithm.

[0115] The microbial-related variable (feature) according to the present disclosure may include the total content and relative abundance of at least one microorganism selected from a first group of microorganisms consisting of microorganisms of the genus Gemella, microorganisms of the genus Streptococcus, microorganisms of the genus Granulicatella, microorganisms of the genus Rothia, microorganisms of the genus Mogibacterium, and microorganisms of the genus Schaalia.

[0116] The microbial-related variable (feature) according to the present disclosure may include the total content and relative abundance of at least one microorganism selected from a second microbial group consisting of microorganisms of the genus Alloprevotella, microorganisms of the genus Veillonella, microorganisms of the genus Haemophilus, microorganisms of the genus Selenomonas, microorganisms of the genus Campylobacter, microorganisms of the genus Cardiobacterium, microorganisms of the genus Megasphaera, and microorganisms of the genus Kingella.

[0117] The learning unit can train a machine learning model using microbial-related variables.

[0118] For example, the learning unit can train a machine learning model to predict the presence and / or type of migraine for each microorganism, particularly for each signature microorganism data, by performing supervised learning based on labels regarding the presence and / or type of migraine for each microorganism data (training data) and the content and variation levels of selected signature microorganisms.

[0119] The machine learning model may include, for example, at least one of a linear regression analysis (LRA) model, a random forest model, a generalized linear model (GLMNET) model, a gradient boosting model, and an extreme gradient boosting (XGB) model. Preferably, the machine learning model may be a random forest model.

[0120] The diagnosis / determination unit can input microbial data collected from the subject to be examined into a trained machine learning model to diagnose and / or determine the presence and / or type of migraine.

[0121] For example, the diagnosis / discrimination unit may diagnose migraine based on the presence and / or type of migraine, which is the output of the machine learning model. In other words, the diagnosis / discrimination unit may determine the presence and / or type of migraine in the subject being examined, or predict the probability and / or severity of migraine in the subject being examined, based on the output of the machine learning model.

[0122] In an exemplary machine learning model according to the present disclosure, optimal modeling for determining the presence and type of migraine was implemented using the total content and relative abundance of specific microorganisms as microbial-related variables (features) (see FIGS. 3a-3c, 9a-9c, and 15a-15f).

[0123] Specifically, first, microorganisms satisfying the prevalence criterion (e.g., less than 10%) among the entire oral microbiota were selected (microorganisms with a prevalence of less than 10%), and features including the selected microbial signatures were trained using a random forest modeling technique to classify migraine in a machine learning model. Through feature importance analysis, it was confirmed that among the oral microbiota, microbial groups that were significantly enriched or decreased in the chronic migraine (CM) or episodic migraine (EM) groups compared to the normal control (HC) group ranked high in migraine classification. Subsequently, the performance of the machine learning model was improved based on a combination of three features: host factors (age, sex, BMI, PC 1-8), microbial factors (oral microbial signature; unlike the previous step where all microbes that met a certain level of prevalence criteria were included, only genus-level microbes that were significantly increased or decreased in common in specific migraine groups (e.g., chronic migraine and episodic migraine groups) were included from the beginning), and these combined. The microbial-based machine learning model constructed in this way enabled analysis of feature importance, and the results of this analysis were reflected in the construction of a more sophisticated machine learning model.

[0124] In the method of the present disclosure, when the relative abundance of at least one microorganism selected from a first microorganism group increases and / or the relative abundance of at least one microorganism selected from a second microorganism group decreases in a plurality of microorganism data extracted from microorganisms collected from a subject to be examined, a diagnosis of migraine can be made.

[0125] Hereinafter, the present disclosure will be described in more detail with reference to examples. It will be apparent to those skilled in the art that these examples are intended solely to illustrate the present disclosure more specifically and that the scope of the present disclosure is not limited by them.

[0126]

[0127] Glossary of terms

[0128] CM = chronic migraine; EM = episodic migraine; HC = healthy (or normal) control group; ns = not significant; PCoA = principal coordinate analysis; AUC = area under the receiver operating characteristic curve; KEGG = Kyoto Encyclopedia of Genes and Genomes; LEfSe = linear discriminant analysis effect size; BMI = body mass index; HIT-6 = Headache Impact Test; MIDAS = Migraine Disability Assessment; LC-MS / MS = liquid chromatography-tandem mass spectrometry; SCFA = short-chain fatty acid; SD = standard deviation; CV = coefficient of variation

[0129]

[0130] [Example 1] Selection and validation of signature microorganisms showing a significant correlation with migraine through oral microbiome analysis alone.

[0131] (1) Participants and clinical evaluation

[0132] Participants were selected to investigate differences in the oral microbiome between migraine patients and healthy controls (HC).

[0133] Migraine patients were selected from among outpatients visiting the Department of Neurology at Severance Hospital with headache who met the diagnostic criteria for migraine as defined in the International Classification of Headache Disorders, Third Revision. Exclusion criteria for participants included: (1) patients under 19 years of age or over 65 years of age; (2) patients with secondary headaches, excluding medication-overuse headaches; (3) patients who refused to participate in the study; (4) patients who were unable to express their opinions due to mental or physical weakness; (5) pregnant or lactating women; (6) patients taking migraine preventive medications; (7) patients who had taken probiotics or antibiotics within the 3 months prior to study participation; (8) patients taking medications for internal diseases or other pre-existing conditions; and (9) other vulnerable study subjects.

[0134] The selected participants were divided into episodic migraine patients and chronic migraine patients, and the following clinical characteristics were investigated according to the migraine diagnostic criteria of the Third Edition of the International Classification of Headache Disorders: headache intensity; headache frequency; headache characteristics including unilateral headache, throbbing headache, and aggravation by activities of daily living (ADL); associated symptoms including nausea, vomiting, photophobia, and phonophobia; Headache Impact Test-6 (HIT6) score; Migraine Disability Assessment score; and aura symptoms including anxiety, depression, and fibromyalgia.

[0135] The normal control (HC) group was selected from people who had not had a headache in the past year and had never experienced a migraine attack in their lifetime.

[0136] The clinical characteristics of migraine patients and the normal control (HC) group are shown in Table 1 below.

[0137]

[0138] CM (N=55) EM (N=55) HC (N=55) p Sex (female) 51 (92.727%) 49 (89.091%) 51 (92.727%) 0.732 Age 45.945 ± 10.78 8 42.182 ± 9.93 9 44.455 ± 11.3 1 30.181 BMI 22.838 ± 3.5 0 22.791 ± 3.3 0 3 23.086 ± 3.6 3 40.892 Monthly migraine frequency 21.055 ± 6.2 4 64.982 ± 2.779 - <0.001 Severity 46 (83.636%) 45 (81.818%) - 1 Unilateral headache 30 (54.545%) 38 (69.091%)-0.17 Pulsating headache54 (98.182%)49 (89.091%)-0.118 ADL46 (83.636%)49 (89.091%)-0.578 Nausea49 (89.091%)50 (90.909%)-1 Vomiting26 (47.273%)33 (60.000%)-0.251 Photophobia36 (65.455%)30 (54.545%)-0.33 Phonophobia41 (74.545%)35 (63.636%)-0.302 HIT6_total63.600 ± 7.72358.036 ± 9.309-0.001 MIDAS_total40.382 ± 46.72815.564 ± 17.018-<0.001 Anxiety 33 (60.000%) 15 (27.273%)-0.001 Depression 33 (60.000%) 15 (27.273%)-0.001 Fibromyalgia 25 (45.455%) 5 (9.091%)-<0.001

[0139]

[0140] (2) Sample collection and DNA extraction

[0141] In Example 1, 1 mL of saliva was collected from each participant in a 50 mL clear saliva container. Participants were instructed not to consume any food, drink, or medication for 30 minutes prior to saliva collection and rinsed their mouths with water immediately prior to saliva collection. The collected saliva was immediately stored in a research freezer at -30°C until analysis.

[0142] The QIAamp DNA kit was used to extract DNA from saliva, and DNA was extracted according to the manufacturer's instructions.

[0143] (3) Confirmation of the type and amount of oral microbiome

[0144] To identify the oral microbiome in the samples extracted in Example 2, the V3-V4 region of 16s rRNA was amplified by PCR. The PCR products were then resolved by 1% agarose gel electrophoresis and visualized using the G: BoxiChemi XL system (Syngene, Cambridge, UK). After attaching the Illumina adapter primer, Nextera® XT Index Kit V2, gene sequencing was performed using the Illumina V3 600 cycle cartridge and Illumina MiSeq equipment (San Diego, California, USA).

[0145] (4) Statistical analysis and results

[0146] The types of oral microbes and the abundance of 16s rRNA were compared between episodic migraine, chronic migraine, and normal control groups using chi-square and ANOVA tests. The correlation between clinical measures of headache frequency, intensity, and headache-related impact in migraine patients and the types and abundance of microbes constituting the oral microbiome were analyzed.

[0147] Perform abundance analysis (species-level, SIAMCAT, p.adj<0.05) and if the relative abundance increased , when relative abundance decreases The results of the analysis are shown in Fig. 1. In chronic migraine and episodic migraine, the relative abundance of Gemella hemolysans, Granulicatella adiasens, Streptococcus gordonii, and Streptococcus pneumoniae was commonly increased compared to the normal control group, and the relative abundance of Centipeda periodontii, Cardiobacterium valbarum, Cattonella morbi, Selenomonas noxia, Haemophilus parainfluenzae, Veillonella artifica, Veillonella hominis, Kingella oralis, Campylobacter concissus, Prevotella aulorum, Veillonella parvula, Megasphaera micronysiformis, and Alloprevotella PAC001345_s was decreased. In episodic migraine, the relative abundance of Porphyromonas pastoris was additionally increased, while the relative abundances of Prevotella oris and Prevotella salibae were decreased. In chronic migraine, the relative abundances of Mosquitobacterium dysversum, Shaalia odontolitica, Shaalia JUPW_s, and Rossia JVPY_s were additionally increased, while the relative abundances of Cardiobacterium hominis, Fusobacterium periodonticum, Neisseria elongata, Prevotella saccharolytica, Veillonella_uc, Prevotella phalens, Prevotella nanseiensis, and Haemophilus parahaemolyticus were decreased. The abundance, significance, and fold change values ​​of each microorganism that showed differences compared to the normal control group are shown in Figure 2.

[0148] A Poisson regression analysis was performed to examine the association between oral microbiome genera that showed significant differences between patients with chronic migraine (CM) and healthy controls (HC), or between patients with episodic migraine (EM) and healthy controls (HC), and monthly migraine frequency. The results are presented in Table 2.

[0149]

[0150] Microorganism EstimateSEz. valuep_valsignifMegaspira0.2350.063.910***Roscia0.0160.0044.3890***Cartonella0.5820.1543.7740***Haemophilus-0.0220.005-4.5850***Shaalia0.0460.0143.1960.001**Mosquitobacterium3.7731.193.170.002**Selenomonas-0.240.097-2.4830.013*Cardiobacterium-0.4590.216-2.1240.034*Allophrenia-0.050.027-1.8750.061. Ranulicatella 0.0220.0131.6640.096. Fusobacterium-0.0150.011-1.4180.156- Porphyromonas-0.0070.005-1.3990.162- Campylobacter-0.0880.064-1.3730.17- Veillonella 0.0060.0 051.2760.202-Jemella-0.0140.017-0.8330.405-Kingella0.1040.3220.3240.746-Streptococcus-0.0010.003-0.2870.774-JQ467121_g-0.1731.048-0.1650.869-

[0151]

[0152] To evaluate the diagnostic performance of migraine using the oral microbiome, receiver operating characteristics (ROC) curve analysis was performed using oral microbe genera that showed significant differences between migraine patients and the healthy control (HC) group. As shown in Fig. 3, the AUC was 0.803 when comparing migraine patients with the healthy control (HC) group (non-headache), 0.776 when comparing chronic migraine (CM) patients with the healthy control (HC) group (non-headache), and 0.789 when comparing episodic migraine (EM) patients with the healthy control (HC) group (non-headache), confirming that the oral microbiome selected through this example has a high diagnostic accuracy for migraine.

[0153]

[0154] [Example 2] Selection and validation of signature microorganisms showing a significant correlation with migraine through simultaneous analysis of the oral and gut microbiomes.

[0155] 1. Method

[0156] (1) Standard protocol approval, registration and patient consent

[0157] Written informed consent was obtained from all participants, and this study was approved by the Institutional Review Board of Severance Hospital (IRB No. 2021-3925-001).

[0158] (2) Participants

[0159] From September 2022 to May 2023, participants with EM and CM were recruited from the outpatient neurology clinic of a tertiary university hospital in Seoul. The inclusion criteria for EM and CM participants were as follows: (1) age 19 to 65 years, and (2) meeting clinical criteria for EM (code 1.1 or 1.2) or CM (code 1.3) of the International Classification of Headache Disorders, Third Revision (ICHD-3). 24

[0160] Participants were excluded if they met the following criteria: (1) medication overuse headache; (2) receipt of preventive treatment for migraine; (3) current or past treatment for oral, gastrointestinal, internal medicine, or psychiatric disorders, excluding anxiety, depression, or fibromyalgia; (4) significant dietary changes in the 3 months prior to the study; (5) use of probiotics or antibiotics in the 6 months prior to the study; (6) pregnancy or lactation.

[0161] Clinical characteristics and common comorbidities of migraine, including anxiety, depression, and fibromyalgia, were assessed in participants with episodic migraine (EM) and chronic migraine (CM).

[0162] In addition, a control group was recruited through advertisements after matching the episodic migraine (EM) group and the chronic migraine (CM) group in terms of age, sex, and body mass index (BMI). The control group was considered eligible for the study if they had not experienced headaches in the past year and had not experienced migraines or migraine attacks in their lifetime. Furthermore, the exclusion criteria (3)-(6) above were applied to the control group. All study participants completed a lifestyle questionnaire, including questions about cohabitation, smoking, sleep patterns, and eating habits.

[0163] (3) Sample size calculation

[0164] The sample size was calculated using the original data set from our previous study [Sci Rep. 2023 Jan 12;13(1):626]. 22 The total required sample size was calculated to be 165, with 55 samples per group.

[0165] To determine the sample size, a power analysis was performed using the alpha (within-sample variability) and beta (between-sample variability) variability from our previous study [Sci Rep. 2023 Jan 12;13(1):626]. The previous study included 42 patients with episodic migraine (EM), 45 patients with chronic migraine (CM), and 43 healthy controls (HC). 1

[0166] First, the Shannon index was used to calculate alpha diversity for all samples. The means and standard deviations of alpha diversity for all samples in the EM, CM, and HC groups were 3.57±0.46, 3.51±0.49, and 3.67±0.37, respectively. Three difference values ​​of 0.3, 0.4, and 0.5 were determined based on the standard deviations of each group, and the sample size was estimated based on the change in the difference value. The effect size for each difference value was calculated by dividing the difference by the average pairwise standard deviation and then taking the average of the resulting values. The effect sizes for difference values ​​of 0.3, 0.4, and 0.5 were 0.68, 0.91, and 1.14, respectively. For a significant difference in alpha diversity to be observed across the three groups, a total of 70 samples (at least 24 samples per group) were required to achieve an effect size of 0.68 (Figure 10).

[0167] Next, the beta diversity of each sample pair was calculated using the Jensen-Shannon distance. The effect size was calculated as follows. The absolute value of the difference in the mean distance values ​​for all groups was calculated pairwise, and the mean was designated as 'A'. The standard deviation of the distances of all samples within each group was calculated and designated as 'S'. In addition, the effect size was calculated by dividing 'A' by the mean pairwise 'S' and then taking the average of the resulting values. The effect size was calculated as 0.54, and the power was calculated based on the effect size and four different α values ​​(0.001, 0.01, 0.05, and 0.1). The power and sample size according to the significance level (α) are shown in Figure 11. Overall, an effect size of 0.54 required 165 samples (at least 55 samples per group), and we observed significant differences in alpha diversity across the three groups at α < 0.05 and power > 80% (Fig. 11).

[0168] Anxiety, Depression, and Fibromyalgia Assessment

[0169] Participants with episodic migraine (EM) and chronic migraine (CM) completed two Korean-validated self-report measures: the Generalized Anxiety Disorder (GAD)-7 to assess anxiety and the Patient Health Questionnaire (PHQ-9) to assess depression. 2,3 Participants were classified as having anxiety and depression if their GAD-7 and PHQ-9 scores were 7 and 9 or higher, respectively. The 2016 American College of Rheumatology diagnostic criteria were used for the diagnosis of fibromyalgia. 4

[0170] (4) Sample collection and 16S rRNA gene sequence analysis

[0171] On the day of consent, unstimulated saliva was collected using 50 mL sterile Eppendorf conical tubes between 9:00 AM and 12:00 PM. Participants were asked to refrain from eating, drinking, chewing, smoking, and brushing their teeth for at least 1 hour before saliva collection. On the same day, blood was collected from the antecubital vein using Becton, Dickinson Vacutainer Serum Separator Tubes (Becton, Dickinson and Company, USA). Blood samples were kept at room temperature for 15 minutes to allow clotting, and then centrifuged at 1500 g for 15 minutes at 4°C to separate the clot and serum. Saliva and serum were immediately stored at -70°C.

[0172] Stool samples were collected within 3 days of consent (or, preferably, as soon as possible) using a stool collection kit (SPL Life Sciences, Korea). Stool samples collected at home were sealed, immediately frozen at -20°C, and delivered to the study site with the provided ice packs within 14 days of collection. They were then stored at -70°C until analysis.

[0173] Targeted sequencing of microbial 16S rRNA genes extracted from saliva and fecal samples was performed.

[0174] DNA extraction, polymerase chain reaction (PCR) amplification, and sequencing were performed according to the procedures described below.

[0175] PCR was performed on DNA extracted from saliva or stool samples collected from participants using specific primers for the 16S rRNA V3-V4 region (fusion primers commonly used to amplify the V3-V4 region of the 16S rRNA gene on the Illumina MiSeq platform: forward 341F primer and reverse 805R primer). The PCR was performed in four steps: denaturation, annealing, extension, and final extension. Initial denaturation was performed at 95°C for 3 min, followed by 25 min of extension. The cycles were denaturation at 95°C for 30 s, primer annealing at 55°C for 30 s, and extension at 72°C for 30 s, followed by a final extension at 72°C for 5 min.

[0176] (5) Measurement of short-chain fatty acids (SCFA)

[0177] Serum acetate, butyrate, and propionate levels were measured using liquid chromatography-tandem mass spectrometry.

[0178] Reagents and chemicals

[0179] Authentic compounds of three straight-chain SCFAs, including acetic (C2), propionic (C3), and butyric (C4) acids, were purchased from Sigma-Aldrich (St. Louis, MO, USA). Analytical reagent grade 3-nitrophenylhydrazine (3NPH) HCl (98%), N-(3-dimethylaminopropyl)-N'-ethylcarbodiimide (EDC) HCl (99%), high-performance liquid chromatography (HPLC) grade pyridine, liquid chromatography-mass spectrometry (LC-MS) grade acetonitrile, water, methanol, and formic acid were also purchased from Sigma-Aldrich. Additionally, [D4]-acetic acid, [D5]-propionic acid, and [D7]-butyric acid were purchased from Cambridge Isotope Laboratories (Tewksbury, MA, USA). Stock solutions of SCFAs were prepared by dissolving them in water, and stock solutions of isotopically labeled compounds were prepared by dissolving them in methanol and stored at -80°C.

[0180] Preparation of standard solutions

[0181] Stock solutions of 1 M acetic acid, 0.1 M butyric acid, and propionic acid were prepared by dissolving them in deionized water. Working solutions with concentrations of 100 mM acetic acid, 10 mM butyric acid, and 10 mM propionic acid were prepared by diluting the stock solutions in deionized water. Stock solutions of 100 μM [D4]-acetic acid, 10 μM [D5]-butyric acid, and [D7]-propionic acid were prepared by dissolving the isotopic internal standards of SCFAs (SCFA-IS) in water. Internal standard working solutions with a concentration of 0.1 mM of each internal calibrator were prepared by diluting the stock internal standard solutions in methanol. Calibrators were prepared by diluting with 2% albumin at six concentrations of acetic acid (5, 10, 25, 50, 100, 250 μM) and seven concentrations of butyric acid and propionic acid (0.2, 0.5, 1, 2.5, 5, 10, 25 μM) and stored at -80°C until use.

[0182] SCFA extraction and derivatization

[0183] SCFAs were analyzed using liquid chromatography-tandem mass spectrometry (LC-MS / MS). Briefly, 10 μL of an aqueous internal standard mixture containing 100 μM [D4]-acetic acid and 10 μM each of [D5]-propionic acid and [D7]-butyric acid was added to 10 μL of serum samples or calibrators. This mixture was prepared according to the method described by Han et al. 5 With some modifications, the method was further applied for 3-nitrophenylhydrazine (3NPH) derivatization. Derivatization was performed by adding 100 μL of 15 mM 3-nitrophenylhydrazine hydrochloride and 2 mM N-(3-dimethylaminopropyl)-N'-ethylcarbodiimide hydrochloride in methanol with 2% pyridine to the sample and internal standard mixture, followed by gentle mixing at 25°C for 30 min. After a centrifugation step (5 min at 16,000 g, repeated three times), 20 μL of the supernatant was transferred to a microvial, and the reaction was stopped by adding 200 μL of 0.1% formic acid. Then, 10 μL of this solution was injected into the LC-MS / MS system.

[0184] LC-MS / MS analysis

[0185] Separation of the analytes was performed using an Agilent 1290 Infinity II LC system (Agilent Technologies Inc., Santa Clara, CA, USA) and an Acquity UPLC BEH C18 column (2.1 mm 100 mm, 1.75 μm particle size, 130Å pore size; Waters, Milford, MA) at a flow rate of 400 μL / min at 40°C. Mobile phase A was deionized water containing 0.01% formic acid, and mobile phase B was acetonitrile containing 0.01% formic acid. The gradient program started with 20% B, held at 20% B for 2 min, then increased to 40% B within 7 min, and then increased to 100% B after 7.5 min and held within 8 min. Finally, it was reduced to 20% B within 8.5 minutes, maintained within 9 minutes, and equilibrated within 3 minutes.

[0186] The analytes were ionized via electrospray ionization in negative ion mode on a QTRAP 5500 triple-quadrupole mass spectrometer (AB SCIEX, Foster City, CA, USA). The analytes were monitored in multiple reaction monitoring mode, and the mass transitions and MS parameters are presented in Table 3.

[0187] Calibration curves were generated by graphically plotting the peak area ratios of the analytes and internal standards over the concentration ranges of 5 to 250 μM (6 points) for acetic acid and 0.2 to 25 μM (7 points) for butyric and propionic acids. Data were collected and analyzed using Analyst 1.6.3 software (AB SCIEX). The intra-assay coefficients of variation (CV) (n = 3), calculated from the measurements of three different concentrations of quality control materials, were 0.38 to 1.42%, 1.33 to 2.05%, and 0.00 to 0.84% ​​for acetic, propionic, and butyric acids, respectively (Tables 4 and 5). The inter-assay coefficients of variation (n= 5) were 6.24 to 15.36%, 3.71 to 7.06%, and 8.21 to 10.80% for acetic acid, propionic acid, and butyric acid, respectively (Tables 4 and 5).

[0188]

[0189] LC-MS / MSCompoundRT (min)Q1 (m / z)Q3 (m / z)DP (V)CE (V)Acetic acid-3NPH3.01194.000152.000-101-19Acetic acid of SCFA-3-nitrophenylhydrazone (3NPH) derivative (D4)-3NPH3.01197.100153.000-130-21Propionic acid-3NPH3.96208.000165.000-75-18Propionic acid (D5)-3NPH3.96212.990153.000-95-22Butyric acid-3NPH5.19222.000152.000-127-22Butyric acid (D7)-3NPH5.19229.004153.100-125-24

[0190]

[0191] The retention times (RT), mass transfers (quantitative and qualifier ions; Q1 and Q3), declustering potentials (DP), and collision energies (CE) of the analytes and stable isotope-labeled internal standards are reported.

[0192]

[0193] Intra-assay and inter-assay precision when analyzed three times over 1 day (intra-assay) or 5 consecutive days (inter-assay) using three concentrations of quality control material. Intra-assay (n= 3)Target (μM)Mean (μM)Bias (%)SD (μM)CV (%)Acetic acid6.05.51-8.170.081.4260.059.39-1.010.230.38300.0307.862.622.180.71Propionic acid0.60.56-6.110.012.056.05.73-4.440.081.4630.029.62-1.270.391.33Butyric acid acid0.60.55-8.330.000.006.05.96-0.720.050.8430.030.632.090.150.48

[0194]

[0195] Intra- and inter-assay precision when analyzed in triplicate over 1 day (intra-assay) or 5 consecutive days (inter-assay) using three concentrations of quality control material. Inter-assay (n= 5)Target (μM)Mean (μM)Bias (%)SD (μM)CV (%)Acetic acid6.05.74-4.350.8815.3660.059.63-0.623.726.24300.0295.94-1.3518.666.31Propionic acid0.60.58-2.630.023.716.05.94-1.070.427.0630.029.06-3.121.936.65Butyric acid acid0.60.600.530.058.686.05.69-5.250.478.2130.029.29-2.373.1610.80

[0196]

[0197] (6) Diversity and taxonomic and functional richness analysis

[0198] Amplicon sequence variant (ASV)-based taxonomic profiling was performed using a pipeline based on internal plugins in QIIME2 (version 2022.11). Taxonomic assignments were performed using a feature classifier plugin referencing the EzBioCloud database (CJ Bioscience Inc., Korea). 25

[0199] For alpha diversity, the number of taxa and Shannon diversity index were calculated, and for beta diversity, the Bray-Curtis dissimilarity was calculated. Differential abundance tests were performed using the SIAMCAT R package to compare the oral and gut microbiota characteristics at the genus level between the episodic migraine (EM) and chronic migraine (CM) groups and the healthy control (HC) group. 26 The analysis included preprocessing options with a prevalence filtering cutoff of 0.1 and multiple testing correction using the Benjamini-Hochberg procedure at a significance level of 0.05. Furthermore, a logistic regression model was constructed to determine whether the identified microbial signatures remained significant after adjusting for potential confounding covariates such as age, sex, body mass index, and lifestyle factors.

[0200] The PICRUSt algorithm was used to estimate the Kyoto Encyclopedia of Genes and Genomes (KEGG) functional pathway profiles based on the taxonomic profiles. 27 Spearman's correlation coefficient was used to investigate the correlation between relative abundance and metabolic pathway features in the oral cavity, which was visualized using a heatmap generated with the Seaborn Python library.

[0201] (7) Correlation between microbial characteristics and clinical characteristics

[0202] To further analyze the clinical significance of the oral microbiome identified in the comparative analysis, Poisson regression analysis was used to assess the association between the relative abundance of oral microbial signatures and the number of headache days per 30 days in migraine patients. Age, sex, body mass index, lifestyle factors, anxiety, depression, and fibromyalgia were adjusted as covariates.

[0203] (8) Correlation network analysis and environmental analysis

[0204] To identify microbial modules exhibiting significant microbial associations, we performed a correlation network analysis of the microbiomes within each body site. Distinct clusters of communities with similar oral and gut microbiota compositions, termed orotypes and enterotypes, were identified using Dirichlet Multinomial Mixtures (DMM) clustering. 28 The optimal number of clusters was determined through model fitting based on the Laplace approximation. Linear discriminant analysis effect size (LEfSe) was used to identify characteristic genera that stratify the microbiome into distinct clusters, using LDA 3.0 as the baseline. 29

[0205] We performed network analysis of the oral and gut microbiomes of the entire sample using Pearson correlation using Python's scipy library, followed by center log-ratio (CLR) transformation and prevalence filtering of taxa at a 10% cutoff point. Adjusted P-values ​​were calculated using the BH procedure using Python's statsmodels library, and cutoffs of absolute rho ≥ 0.4 and adjusted P < 0.05 were used as indicators of significant network associations. To identify clusters of taxa with positive associations in each body site, we applied the Louvain clustering algorithm using Python's community library. The network was visualized using Cytoscape 3.9.1, with node size and color annotated with the average abundance percentage and the number of participants sharing the same amplicon sequence variant in the oral and gut microbiomes of each genus. For the oral microbiome, node labels were highlighted based on the abundance difference results.

[0206] (9) Oral-intestinal transmission analysis

[0207] We analyzed the possibility of oral-gut transmission using ASVs shared between the oral cavity and the gut. ASVs identified in both the oral and gut microbiomes of each individual were classified as shared. Otherwise, they were classified as present only in either the oral or gut microbiome. The average abundance of shared oral and gut microbiomes was then compared between chronic migraine (CM), episodic migraine (EM), and healthy control (HC) groups. Furthermore, we investigated the presence of oral microbial signatures in the gut in each group.

[0208] (10) Principal coordinate analysis (PCA) of lifestyle habit survey data

[0209] To identify and adjust for potential confounding factors that may influence the relationship between the microbiome and migraine, we conducted a self-reported lifestyle survey. Categorical variables derived from the survey were converted to dummy variables, and ordered categories were treated as numeric integer variables. Preprocessed variables were scaled, and PCA (n_components=25, svd_solver='full') was performed using the sklearn Python library. Eight principal component (PC) axes were selected because they exceeded the cutoff point after evaluating a scree plot, which compares the proportion of explained variance of each PC axis to the mean. Furthermore, the top 20 PC loadings for the eight selected PC axes were evaluated, and the coordinates of each axis were compared across groups (Figures 12a-12d).

[0210] (11) Statistical analysis

[0211] Statistical analysis was performed using the moonbook R package with default options. Continuous variables were compared using independent t-tests or analysis of variance for normally distributed data, and Wilcoxon rank-sum tests or Kruskal-Wallis tests for nonparametric data. Categorical data were compared using Pearson's χ² test or Fisher's exact test. Principal components analysis was performed for lifestyle questionnaire data using the sklearn Python library. Finally, eight principal component (PC) axes were selected and used as covariates in modeling to adjust for potential confounders in the relationship between the microbiome and migraine.

[0212] 2. Results

[0213] (1) Characteristics of patients with episodic migraine (EM) and chronic migraine (CM) and healthy (normal) controls (HC)

[0214] A total of 165 participants were enrolled, 55 each in the episodic migraine (EM), chronic migraine (CM), and healthy control (HC) groups. Demographic and clinical characteristics are presented in Table 6 . There were no significant differences in gender, age, or BMI among the three groups. However, the number of headache days per 30 days, the Headache Impact Test-6 (HIT-6) total score, the Migraine Disability Assessment Scale (MIDAS) total score, and the prevalence of anxiety, depression, and fibromyalgia were significantly different between the EM and CM groups (Figures 13a-13i).

[0215]

[0216] Demographic and clinical characteristics of participants CM (n= 55) EM (n= 55) HC (n= 55) p value Sex, female, n (%) 51 (92.7%) 49 (89.1%) 51 (92·7%) 0.73 2 Age, years 45.9 ± 10.8 42.2 ± 9.9 44.5 ± 11.3 0.18 1 BMI, kg / m 2 22.8 ± 3.522.8 ± 3.323.1 ± 3.60.892Number of headache days per 30 days21.1 ± 6.25.0 ± 2.8-< 0.001 *evere pain intensity,nn(%)46 (83.6%)45 (81.8%)-> 0.999Unilateral location,n(%)30 (54.5%)38 (69.1%)-0.170Pulsating quality,n(%)54 (98.2%)49 (89.1%)-0.118Aggravation by routine physical activity,n(%)46 (83.6%)49 (89.1%)-0.578Nausea,n(%)49 (89.1%)50 (90.9%)-> 0.999Vomiting,n(%)26 (47.3%)33 (60.0%)-0.251Photophobia,n(%)36 (65.5%)30 (54.5%)-0.330Phonophobia,n(%)41 (74.5%)35 (63.6%)-0.302Total HIT-6 score63.6 ± 7.758.0 ± 9.3-0.001 * Total MIDAS score40.4 ± 46.715.6 ± 17.0-< 0.001 * Anxiety,n(%)33 (60.0%)15 (27.3%)-0.001 * Depression,n(%)33 (60.0%)15 (27.3%)-0.001 * Fibromyalgia,n(%)25 (45.5%)5 (9.1%)-< 0.001 *

[0217] Age, sex, and BMI were compared between the chronic migraine (CM), episodic migraine (EM), and healthy control (HC) groups. Other clinical variables were compared between the chronic migraine (CM) and episodic migraine (EM) groups. *p<0.05.

[0218]

[0219] (2) Overall composition and microbial diversity

[0220] The overall composition of the oral and gut microbiomes of the three groups was investigated at the genus level, respectively (Figs. 4a and 4b). Alpha diversity did not show a significant difference at the genus level in the oral microbiome (Fig. 4a). However, for the gut microbiome, the chronic migraine (CM) group had a higher Shannon diversity than the episodic migraine (EM) group (p = 0.027), and the number of observed features was higher than that in the episodic migraine (EM) (p = 0.034) and healthy control (HC) (p = 0.040) groups (Fig. 4b). Analysis of genus-level beta diversity of the oral and gut microbiomes revealed significant differences among the three groups (oral and gut microbiome, p = 0.001 and 0.039, respectively). The differences were particularly pronounced in the oral microbiome (data not shown).

[0221] (3) Taxonomic and functional characteristics of the oral microbiome

[0222] In the chronic migraine (CM) group and the episodic migraine (EM) group, 17 and 15 oral microbial signatures were identified at the genus level, respectively, which were significantly enriched or decreased compared to the HC group (Figs. 5a and 5b). In both the chronic migraine (CM) and episodic migraine (EM) groups, the abundance of Gemella, Streptococcus, Granulicatella, and Mogibacterium was significantly increased, while the abundance of Alloprevotella, Veillonella, Haemophilus, Selenomonas, Catonella, Campylobacter, Cardiobacterium, Megasphaera, and Kingella was significantly decreased (Fig. 5c). In addition, the chronic migraine (CM) group showed an enrichment of Schaalia and Rothia. However, the relative abundance of the gut microbiome did not significantly differ between the chronic migraine (CM) and episodic migraine (EM) groups compared to the normal control (HC) group. After adjusting for age, sex, body mass index, and lifestyle factors, the statistical significance of the oral microbiome remained largely intact after applying a logistic regression model (Table 7). Furthermore, compared to the normal control (HC) group after adjustment, the episodic migraine (EM) group showed a significant enrichment of Rothia.

[0223] A total of 62 and 52 KEGG metabolic pathways were identified as significant functional markers of the oral microbiome in the chronic migraine (CM) and episodic migraine (EM) groups, respectively. Of these, 48 were identified as overlapping markers (Figures 6a and 6b). Carbohydrate metabolism, including fructose / mannose, galactose, starch / sucrose, and glycolysis / gluconeogenesis, was enhanced in both the chronic migraine (CM) and episodic migraine (EM) groups, as well as glycan biosynthesis and metabolism. In contrast, butanoic acid, propanoic acid, and nitrogen metabolism pathways were decreased in the migraine group.

[0224]

[0225] Logistic regression analysis of oral microbiota signatures showing significant differences between chronic migraine (CM) and healthy control (HC) groups or episodic migraine (EM) and healthy control (HC) groups. ComparisonFeatureEstimateSEz valueP-valueSignificanceEMVeillonella-0.3470.08-4.354<0.001***EMStreptococcus0.1060.0273.86<0.001***EMCatonella-10.733.025-3.548<0.001***CMVeillonella-0.1420.04-3.577<0.001***CMStreptococcus0.1380.0353.976<0.001*** EMCampylobacter-1.9250.496-3.885<0.001***CMRothia0.1380.0373.701<0.001***CMCampylobacter-1.5820.495-3.1970. 001**EMGemella1.0260.2993.4320.001***CMGemella0.9570.2873.3380.001***EMMegasphaera-2.0710.671-3.0890.002**C MSchaalia0.8220.2723.0250.002**CMHaemophilus-0.1380.044-3.1560.002**EMPorphyromonas0.1940.0652.9930.003**CM Fusobacterium-0.3210.107-3.0040.003**CMGranulicatella0.3120.1072.9090.004**CMCardiobacterium-5.5952.057-2.7 20.007**CMSelenomonas-2.1350.825-2.5870.010**EMKingella-6.5012.576-2.5240.012*EMSelenomonas-1.7090.692-2.47 10.013*CMKingella-6.2932.563-2.4550.014*EMHaemophilus-0.0820.035-2.310.021*CMAY093465_g-27.21711.754-2.3150.021*CMAlloprevotella-0.5820.26-2.2390.025*EMGranulicatella0.2270.1042.1790.029*EMRothia0.0880.0412.160.031*EMCardiobacterium-2.7661.32 3-2.090.037*CMMegasphaera-0.830.416-1.9940.046*EMAlloprevotella-0.4340.22-1.9730.049*EMMogibacterium61.36432.4121.8930.058-CMMogibacter ium57.60632.8361.7540.079-EMSchaalia0.4180.2411.7340.083-EMFusobacterium-0.1340.078-1.7270.084-CMPorphyromonas0.1010.0641.590.112-CMCat onella-2.6781.755-1.5260.127-EMAY093465_g-9.1196.448-1.4140.157-EMJ Q467121_g-8.1117.326-1.1070.268-CMJQ467121_g-3.4045.205-0.6540.513-.

[0226] [Note] Age, sex, BMI, and lifestyle factors principal components (PCs 1-8) were adjusted as covariates for the stratified chronic migraine (CM) vs. healthy controls (HC) or episodic migraine (EM) vs. healthy controls (HC) data sets. ***P< 0.001; **P< 0.01; *P< 0.05; ns, not significant.

[0227]

[0228] Specifically, to illustrate the association between taxonomic oral signatures and inferred functional KEGG metabolic pathway signatures, a heatmap of Spearman correlation results for the percent abundance of pathways and taxa was generated, revealing significant correlations and clustering between the abundances of oral microbiota and metabolic signatures. Specifically, the genera Streptococcus, Granulicatella, Gemella, and Schaalia showed positive correlations with pathways related to sugar metabolism, whereas the genera Veillonella, Selenomonas, and Campylobacter showed positive correlations with SCFA and nitrogen-related metabolic pathways (data not shown).

[0229] (4) Association between oral microbiota characteristics and monthly headache days

[0230] The relative abundance of Rothia, Granulicatella, Schaalia, Megasphaera, Catonella, and Mogibacterium showed a significant positive correlation with the number of headache days per 30 days, whereas Cardiobacterium, Selenomonas, Alloprevotella, and Haemophilus showed a negative correlation (Table 8 and Figures 13a-13i).

[0231]

[0232] Poisson regression analysis of the association between the number of headache days per 30 days in the migraine group and significantly enriched or reduced oral microbial signatures compared to the control group. Oral microbial signatures at genus level. EstimateStandard Error. z value. p value. Significance. Rothia. 0.017. 0.004. 4.445. < 0·001. ***Catonella. 0.707. 0.154. 4.587. < 0·001. ***Haemophilus. 0.022. 0.005. 4.487. < 0·001***Megasphaera0.1960.063.2520.001**Mogibacterium3.9211.2073.2490.001**Schaalia0.0370.0142.5990.009**Granulicatella0.0330.0142.4 650.014*Alloprevotella-0.0570.026-2.1550.031*Selenomonas-0.2020.094-2.150.032*Cardiobacterium-0.4390.21-2.0860.037*AY093465_g-2.8791. 487-1.9360.053nsFusobacterium-0.020.011-1.8620.063nsKingella0.4010.3231.2410.215nsPorphyromonas-0.0040.005-0.6890.491nsJQ467121_g0.7 051.0590.6650.506nsGemella-0.0080.017-0.5010.616nsCampylobacter-0.0280.065-0.4230.673nsVeillonella0.0010.0050.250.803nsStreptococcus< 0·0010.0030.0150.988ns

[0233] [Note] Age, sex, body mass index (BMI), lifestyle factors (PC 1-8), anxiety, depression, and fibromyalgia were adjusted as covariates. ***p< 0.001; **p< 0.01; *p< 0.05.

[0234]

[0235] (5) Serum SCFA levels

[0236] In the chronic migraine (CM) group, serum propionic acid levels were significantly reduced compared to the episodic migraine (EM) group (p < 0.001) and the healthy control (HC) group (p < 0.001). However, no significant differences were observed in serum acetic acid and butyric acid levels among the three groups (Figs. 14a-14c).

[0237] (6) Oral and intestinal microbial networks and community levels

[0238] Correlation network analysis identified several distinct microbial communities at each sample site (Figures 9a-9f). Among the genera enriched in migraine, Rothia, Gemella, Streptococcus, and Granulicatella were clustered together. Furthermore, distinct subnetwork clusters were identified within the gut taxonomic group, and among these, a notable gut microbiome representing the oral microbial signature for migraine was identified, including Rothia, Gemella, Streptococcus, Granulicatella, and Veillonella.

[0239] Meanwhile, the overall ecological structure of the oral cavity and intestine was evaluated using the DMM clustering method (Figs. 7a-7f). The three oral clusters exhibited different microbial profiles. Cluster E1 was more abundant in Rothia, Gemella, Streptococcus, and Granulicatella. Cluster E2 was more abundant in Neisseria, Haemophilus, Porphyromonas, and Fusobacterium. Cluster E3 was more abundant in Prevotella, Veillonella, Schaalia, and Megasphaera. Significant differences in the prevalence of oral clusters were observed between the chronic migraine (CM), episodic migraine (EM), and healthy control (HC) groups, and the HC group had a significantly lower E1 rate than the chronic migraine (CM) and episodic migraine (EM) groups (p < 0.001) (Figs. 7a-7c).

[0240] Two distinct clusters were identified for the gut microbiome. Cluster E1 contained more Prevotella, Bacteroides, Alistipes, and Phocaeicola, whereas cluster E2 contained more Blautia, Bifidobacterium, and Escherichia. However, the proportions of the two clusters did not differ significantly across the three groups (p = 0.379) (Figs. 7d-7f).

[0241] (7) Oral-intestinal transmission analysis

[0242] The mean abundance percentage of shared ASVs did not differ significantly among the three groups (p = 0.147 and p = 0.133 for oral and intestinal, respectively). The eight most commonly coexisting taxa were Streptococcus, Rothia, Veillonella, Haemophilus, Granulicatella, Schaalia, Gemella, and Megasphaera (Tables 9 and 10).

[0243]

[0244] Genus-level occurrence of signature genera in the oral and gut microbiome and participant-level occurrence of common or distinct ASVs. Enterobacteriaceae Oral CMEMHCCMEMHCAY093465_g0.0 (0.0%)0.0 (0.0%)0.0 (0.0%)6.0 (10.9%)14.0 (25.5%)16.0 (29.1%)Mogibacterium0.0 (0.0%)0.0 (0.0%)0.0 (0.0%)14.0 (25.5%)11.0 (20.0%)2.0 (3.6%)JQ467121_g0.0 (0.0%)0.0 (0.0%)0.0 (0.0%)8.0 (14.5%)5.0 (9.1%)17.0 (30.9%)Selenomonas0.0 (0.0%)0.0 (0.0%)0.0 (0.0%)48.0 (87.3%)50.0 (90.9%)54.0 (98.2%)Cardiobacterium0.0 (0.0%)0.0 (0.0%)0.0 (0.0%)45.0 (81.8%)44.0 (80.0%)50.0 (90.9%)Haemophilus17.0 (30.9%)19.0 (34.5%)17.0 (30.9%)55.0 (100.0%)55.0 (100.0%)55.0 (100.0%)Megasphaera7.0 (12.7%)8.0 (14.5%)14.0 (25.5%)41.0 (74.5%)41.0 (74.5%)52.0 (94.5%)Porphyromonas6.0 (10.9%)8.0 (14.5%)9.0 (16.4%)54.0 (98.2%)54.0 (98.2%)55.0 (100.0%)Catonella0.0 (0.0%)0.0 (0.0%)0.0 (0.0%)37.0 (67.3%)31.0 (56.4%)46.0 (83.6%)Kingella0.0 (0.0%)0.0 (0.0%)0.0 (0.0%)31.0 (56.4%)30.0 (54.5%)47.0 (85.5%)Gemella5.0 (9.1%)9.0 (16.4%)7.0 (12.7%)55.0 (100.0%)55.0 (100.0%)55.0 (100.0%)Granulicatella8.0 (14.5%)10.0 (18.2%)11.0 (20.0%)54.0 (98.2%)55.0 (100.0%)55.0 (100.0%)Fusobacterium10.0 (18.2%)9.0 (16.4%)9.0 (16.4%)55.0 (100.0%)55.0 (100.0%)55.0 (100.0%)Streptococcus54.0 (98.2%)51.0 (92.7%)49.0 (89.1%)55.0 (100.0%)55.0 (100.0%)55.0 (100.0%)Alloprevotella5.0 (9.1%)6.0 (10.9%)2.0 (3.6%)49.0 (89.1%)51.0 (92.7%)55.0 (100.0%)Rothia23.0 (41.8%)13.0 (23.6%)8.0 (14.5%)55.0 (100.0%)55.0 (100.0%)55.0 (100.0%)Campylobacter0.0 (0.0%)1.0 (1.8%)2.0 (3.6%)54.0 (98.2%)54.0 (98.2%)55.0 (100.0%)Schaalia8.0 (14.5%)11.0 (20.0%)9.0 (16.4%)53.0 (96.4%)54.0 (98.2%)54.0 (98.2%)Veillonella19.0 (34.5%)24.0 (43.6%)25.0 (45.5%)55.0 (100.0%)55.0 (100.0%)55.0 (100.0%).

[0245] [Note] Amplicon sequence variant (ASV); chronic migraine (CM); episodic migraine (EM); healthy control (HC)

[0246]

[0247] Genus-level occurrence of signature genera in the oral and gut microbiome and number of participants with common or distinct ASVs Number of participants with common or distinct ASVs No ASVs ASVs in the oral or gut ASVs shared between the oral and gut CMEMHCCMEMHCCMEMHCAY093465_g49.0 (89.1%)41.0 (74.5%)39.0 (70.9%)6.0 (10.91%)14.0 (25.5%)16.0 (29.1%)0.0 (0.0%)0.0 (0.0%)0.0 (0.0%)Mogibacterium 41.0 (74.5%)44.0 (80.0%)53.0 (96.4%)14.0 (25.5%)11.0 (20.0%)2.0 (3.6%)0.0 (0.0%)0.0 (0.0%)0.0 (0.0%)JQ467121_g47.0 (85.5%)50.0 (90.9%)38.0 (69.1%)8.0 (14.5%)5.0 (9.1%)17.0 (30.9%)0.0 (0.0%)0.0 (0.0%)0.0 (0.0%)Selenomonas7.0 (12.7%)5.0 (9.1%)1.0 (1.8%)48.0 (87.3%)50.0 (90.9%)54.0 (98.2%)0.0 (0.0%)0.0 (0.0%)0.0 (0.0%)Cardiobacterium10.0 (18.2%)11.0 (20.0%)5.0 (9.1%)45.0 (81.8%)44.0 (80.0%)50.0 (90.9%)0.0 (0.0%)0.0 (0.0%)0.0 (0.0%)Haemophilus0.0 (0.0%)0.0 (0.0%)0.0 (0.0%)46.0 (83.6%)38.0 (69.1%)40.0 (72.7%)9.0 (16.4%)17.0 (30.9%)15.0 (27.3%)Megasphaera12.0 (21.8%)12.0 (21.8%)2.0 (3.6%)43.0 (78.2%)40.0 (72.7%)52.0 (94.5%)0.0 (0.0%)3.0 (5.5%)1.0 (1.8%)Porphyromonas1.0 (1.8%)1.0 (1.8%)0.0 (0.0%)54.0 (98.2%)54.0 (98.2%)55.0 (100.0%)0.0 (0.0%)0.0 (0.0%)0.0 (0.0%)Catonella18.0 (32.7%)24.0 (43.6%)9.0 (16.4%)37.0 (67.3%)31.0 (56.4%)46.0 (83.6%)0.0 (0.0%)0.0 (0.0%)0.0 (0.0%)Kingella24.0 (43.6%)25.0 (45.5%)8.0 (14.5%)31.0 (56.4%)30.0 (54.5%)47.0 (85.5%)0.0 (0.0%)0.0 (0.0%)0.0 (0.0%)Gemella0.0 (0.0%)0.0 (0.0%)0.0 (0.0%)54.0 (98.2%)49.0 (89.1%)53.0 (96.4%)1.0 (1.8%)6.0 (10.9%)2.0 (3.6%)Granulicatella1.0 (1.8%)0.0 (0.0%)0.0 (0.0%)51.0 (92.7%)48.0 (87.3%)51.0 (92.7%)3.0 (5.5%)7.0 (12.7%)4.0 (7.3%)Fusobacterium0.0 (0.0%)0.0 (0.0%)0.0 (0.0%)55.0 (100.0%)55.0 (100.0%)55.0 (100.0%)0.0 (0.0%)0.0 (0.0%)0.0 (0.0%)Streptococcus0.0 (0.0%)0.0 (0.0%)0.0 (0.0%)5.0 (9.1%)10.0 (18.2%)11.0 (20.0%)50.0 (90.9%)45.0 (81.8%)44.0 (80.0%)Alloprevotella6.0 (10.9%)4.0 (7.3%)0.0 (0.0%)49.0 (89.1%)51.0 (92.7%)55.0 (100.0%)0.0 (0.0%)0.0 (0.0%)0.0 (0.0%)Rothia0.0 (0.0%)0.0 (0.0%)0.0 (0.0%)36.0 (65.5%)45.0 (81.8%)48.0 (87.3%)19.0 (34.5%)10.0 (18.2%)7.0 (12.7%)Campylobacter1.0 (1.8%)1.0 (1.8%)0.0 (0.0%)54.0 (98.2%)54.0 (98.2%)55.0 (100.0%)0.0 (0.0%)0.0 (0.0%)0.0 (0.0%)Schaalia2.0 (3.6%)0.0 (0.0%)1.0 (1.8%)51.0 (92.7%)52.0 (94.5%)51.0 (92.7%)2.0 (3.6%)3.0 (5.5%)3.0 (5.5%)Veillonella0.0 (0.0%)0.0 (0.0%)0.0 (0.0%)42.0 (76.4%)43.0 (78.2%)38.0 (69.1%)13.0 (23.6%)12.0 (21.8%)17.0 (30.9%).

[0248] [Note] Amplicon sequence variant (ASV); chronic migraine (CM); episodic migraine (EM); healthy control (HC)

[0249]

[0250] [Example 6] Building a machine learning model for classifying the presence and type of migraine based on oral microbiome data.

[0251] A machine learning model for migraine classification was built based on a random forest model (chronic migraine (CM) or episodic migraine (EM)), and an AUC of 0.84 ± 0.06 was obtained using microbial features (i.e., oral microbial features of migraine), and an AUC of 0.85 ± 0.05 was obtained using a combination of host and microbial features (Fig. 9a).

[0252] For chronic migraine (CM) classification, the best performance was achieved using both host and microbial features with an AUC of 0.89 ± 0.07, followed by microbial features with an AUC of 0.88 ± 0.07 (Fig. 9b).

[0253] For episodic migraine (EM) classification, the model achieved an AUC of 0.85 ± 0.08 using both host and microbial features, and an AUC of 0.83 ± 0.09 using microbial features (Fig. 9c). Feature importance values ​​for Veillonella and Alloprevotella were high in all three migraine classifications (Figs. 9d-9f).

[0254] Consideration

[0255] We simultaneously investigated the oral and gut microbiomes of participants with episodic migraine (EM), chronic migraine (CM), and healthy control (HC) groups. Key findings include: (1) significant changes in the oral microbiome were observed in the migraine group compared to the healthy control (HC) group, but no significant differences in the gut microbiome. (2) increased relative abundances of Gemella, Streptococcus, Granulicatella, and Rothia, and decreased relative abundances of Alloprevotella, Veillonella, Haemophilus, Selenomonas, Campylobacter, Cardiobacterium, Megasphaera, and Kingella were observed in both the episodic migraine (EM) and chronic migraine (CM) groups, and these differences remained statistically significant after adjusting for demographic and lifestyle factors. (3) the relative abundances of some of these oral genera were significantly associated with monthly headache days after adjusting for demographic and lifestyle factors and psychiatric disorders. (4) Within the migraine group, enriched oral microbiota were associated with carbohydrate metabolism pathways, whereas depleted oral microbiota were associated with SCFA and nitrogen-related pathways. (5) Oral microbiome characteristics of migraine were observed simultaneously in the oral cavity and gut, forming several distinct clusters. (6) Machine learning discrimination using the oral microbiome effectively discriminated migraine status with a high AUC of 0.83–0.88.

[0256] In this disclosure, migraine patients were classified into episodic migraine (EM) (basilar headache) and chronic migraine (CM) (severe headache) groups based on clinical and biological distinctions currently accepted in headache research. According to ICHD-3, chronic migraine (CM) is defined as headache occurring on at least 15 days per month for at least 3 months, with migraine-like symptoms present on at least 8 days. 24 Chronic migraine (CM) is increasingly recognized as a distinct clinical entity characterized by, in addition to frequency, increased disease burden, increased psychiatric comorbidity, and increased risk of medication overuse. 30, 31

[0257] Moreover, chronic migraine (CM) is accompanied by pathophysiological changes, including increased cortical excitability, persistent neuroinflammation, and increased central sensitization, which contribute to impaired pain modulation compared to episodic migraine (EM). 32 These differences provided the rationale for examining episodic migraine (EM) and chronic migraine (CM) separately in this study. Recent studies have shown that patients with chronic migraine (CM), particularly those overdosed on medications, exhibit increased intestinal permeability and systemic inflammation, which may disrupt the gut-brain axis and promote the chronicity of migraine. 33

[0258] Despite numerous studies demonstrating an imbalance between oral and gut microbiota in several neurological disorders, 13-15 Research on the oral microbiome in relation to migraine is lacking. 21, 34 In contrast, previous studies have reported that oral health conditions such as periodontitis can directly or indirectly worsen migraines and increase the risk of chronicity. 35, 36 We hypothesized that there is a link between the imbalance between oral and gut microbiota and migraine, and identified unique taxonomic and functional characteristics of the oral microbiome for migraine.

[0259] Given the anatomical proximity of the trigeminal vascular system to the oral cavity, significant changes in oral health, including local inflammation and oral gut microbiota imbalance, may directly or indirectly influence the onset of migraine. Furthermore, previous studies suggest that saliva may provide valuable insights into migraine, highlighting several biomarkers in saliva, such as enzymes, hormones, and neuropeptides. 5-11 The results of this study suggest that the salivary microbiome may serve as another promising biomarker for migraine.

[0260] Moreover, the oral microbiome is known to play a crucial role in the first step of the nitrate-nitrite-nitric oxide pathway. 37 Nitrates derived from dietary or endogenous metabolism accumulate in saliva and are converted to nitrite by oral microorganisms. Nitric oxide is known to be involved in the pathophysiology of migraine by inducing cerebral vasodilation, neurogenic inflammation, and affecting pain processing in the central nervous system. 38 The abundance of the genera Gemella, Streptococcus, and Granulicatella was significantly increased in both the chronic migraine (CM) and episodic migraine (EM) groups, and the abundance of the genera Rothia and Schaalia was increased in the chronic migraine (CM) group compared to the normal control (HC) group.

[0261] Previous studies 34Similarly, all of these genera contain species capable of reducing nitrate and / or nitrite. However, some genera (Veillonella, Haemophilus, Selenomonas, Kingella) that were significantly depleted in both the chronic migraine (CM) and episodic migraine (EM) groups also contained nitrate-reducing species. These results are similar to those observed in previous studies, which showed that dietary nitrate increased the relative abundance of certain nitrate-reducing microorganisms (e.g., Rothia) while simultaneously decreasing the abundance of others (e.g., Veillonella). 39 Further research using longitudinal design is needed to clarify the implications of these results.

[0262] In this study, we identified several distinct genus clusters within the oral microbiome, with a significantly higher prevalence of clusters comprised of Rothia, Gemella, Streptococcus, and Granulicatella in migraine patients. Furthermore, these oral microbes form colonies within the gut, suggesting that they may possess unique characteristics or niche habits within the microbial community. Oral microbes are generally known to exhibit limited colonization in healthy guts.

[0263] However, in pathological conditions such as inflammatory bowel disease, oral pathogens are more likely to be abundant in the gut, further strengthening the link between the oral and gut microbiota. 19 Although the number and abundance of taxa shared between the oral and gut microbiota were low, rare but selective taxa, including these four oral genera, were identified in the gut microbiota. These observations highlight the importance of these oral microbes, given their abundance in disease groups and their potential as diagnostic and therapeutic targets.

[0264] The present disclosure has the following advantages: (1) The present disclosure is different from previous studies 22 (2) This study simultaneously examined the oral and gut microbiomes of individuals with migraine. (3) This study investigated and adjusted for confounding factors, including demographic, lifestyle, and psychiatric factors known to contribute to variability across microbiome studies. Anxiety, depression, and fibromyalgia are particularly common in people with migraine. 30, 40 Independent associations have been reported between these symptoms and changes in the oral and gut microbiota. 41-43 In the cohort used, anxiety, depression, and fibromyalgia were reported in 27%, 27%, and 9% of episodic migraine (EM) cases, respectively, and in 60%, 60%, and 46% of chronic migraine (CM) cases, respectively. Recognizing the potential for confounding, these conditions were carefully assessed and adjusted for in the analyses. While subgroup analyses excluding participants with comorbid conditions might have provided additional insights, the limited number of eligible participants precluded such analyses without compromising statistical validity. Furthermore, given the high prevalence of these conditions in migraineurs, the findings from the cohort without comorbid conditions may not generalize to the broader migraine population. Nevertheless, the observed changes in the oral microbiome remained consistent after adjustment, supporting the importance of oral microbiome signatures in relation to migraine disease status.

[0265] In conclusion, our findings suggest that oral gut microbiota dysbiosis may play a role in the pathogenesis of migraine, highlighting specific oral microbial communities as potential diagnostic biomarkers and therapeutic targets for migraine.

[0266] The method for providing information for diagnosing and identifying migraines and the machine learning model according to the present disclosure can quickly and accurately diagnose and identify migraines without causing pain or discomfort to the subject (patient) in a non-invasive manner, and the oral signature microbial information according to the present disclosure enables specific diagnosis and identification by distinguishing between migraines, particularly episodic migraines and chronic migraines, and can determine the severity of migraines, making it useful for personalized migraine diagnosis and treatment. In addition, the oral signature microbial(s) identified through the present disclosure can be utilized as biomarkers for diagnosing and / or identifying migraines, and are highly likely to be utilized as promising targets for research or product development in related technical fields, including pharmaceuticals, health functional foods, etc., for use in improving and / or treating migraines and related diseases in the future, and therefore, are expected to have great utility in related industrial fields.

[0267]

[0268] References:

[0269]

[0270]

[0271]

[0272]

[0273]

[0274]

Claims

A method for providing information for diagnosing migraine, comprising the step of measuring the relative abundance of at least one microorganism selected from the group consisting of the genera Megasphaera, Catonella, Haemophilus and Selenomonas in a biological sample obtained from an individual. A method for providing information for diagnosing migraine, further comprising the step of measuring the relative abundance of at least one microorganism selected from the group consisting of the genus Gemella, the genus Granulicatella, and the genus Streptococcus in the biological sample according to claim 1. A method for providing information for diagnosing migraine according to claim 1, wherein a diagnosis of migraine is made when the relative abundance of at least one microorganism selected from the group consisting of the genus Megasphaera, the genus Catonella, the genus Haemophilus, and the genus Selenomonas is decreased compared to a normal control group. A method for providing information for diagnosing migraine according to claim 2, wherein a diagnosis of migraine is made when the relative abundance of at least one microorganism selected from the group consisting of the genus Gemela, the genus Granulocatella, and the genus Streptococcus is increased compared to a normal control group. A method for providing information for diagnosing migraine according to claim 1, further comprising a step of measuring the relative abundance of at least one microorganism selected from the group consisting of the genus Veillonella, the genus Centipeda, the genus Kingella, the genus Campylobacter, the genus Prevotella, the genus Cardiobacterium, and the genus Alloprevotella in the biological sample. A method for providing information for diagnosing migraine according to claim 5, wherein a diagnosis of migraine is made when the relative abundance of at least one microorganism selected from the group consisting of the genus Veillonella, the genus Centipeda, the genus Kingella, the genus Campylobacter, the genus Prevotella, the genus Cardiobacterium, and the genus Alloprevotella is decreased compared to a normal control group. A method for providing information for diagnosing migraine according to claim 1, wherein the biological sample is saliva. A method for providing information for diagnosing migraine according to claim 5, wherein episodic migraine (EM) is diagnosed when the relative abundance of microorganisms of the genus Prevotella is reduced compared to a normal control group, and the microorganisms of the genus Prevotella include at least one selected from the group consisting of Prevotella oris and Prevotella salivae. A method for providing information for diagnosing migraine, further comprising the step of measuring the relative abundance of microorganisms of the genus Porphyromonas in the biological sample according to claim 1. A method for providing information for diagnosing migraine according to claim 9, wherein episodic migraine is diagnosed when the relative abundance of microorganisms of the genus Porphyromonas is increased compared to a normal control group. A method for providing information for diagnosing migraine according to claim 5, wherein chronic migraine (CM) is diagnosed when the relative abundance of microorganisms of the genus Prevotella is reduced compared to a normal control group, and the microorganisms of the genus Prevotella include at least one selected from the group consisting of Prevotella saccharolytica, Prevotella pallens, and Prevotella nanceiensis. A method for providing information for diagnosing migraine according to claim 5, wherein the relative abundance of microorganisms of the genus Cardiobacterium is reduced compared to a normal control group, and when the microorganism of the genus Cardiobacterium is Cardiobacterium hominis, chronic migraine is diagnosed. A method for providing information for diagnosing migraine, further comprising the step of measuring the relative abundance of at least one microorganism selected from the group consisting of the genus Rothia, the genus Mogibacterium, and the genus Schaalia in the biological sample. A method for providing information for diagnosing migraine according to claim 13, wherein chronic migraine is diagnosed when the relative abundance of at least one microorganism selected from the group consisting of the genus Rossia, the genus Mosquitobacterium, and the genus Shaalia is increased compared to a normal control group. A method for providing information for diagnosing migraine, further comprising the step of measuring the relative abundance of at least one microorganism selected from the group consisting of the genus Fusobacterium, the genus Haemophilus, and the genus Neisseria in the biological sample according to claim 1. A method for providing information for diagnosing migraine according to claim 15, wherein chronic migraine is diagnosed when the relative abundance of at least one microorganism selected from the group consisting of the genus Fusobacterium, the genus Haemophilus, and the genus Neisseria is decreased. A method for diagnosing and / or determining migraine using a machine learning model performed on a device, A step of extracting multiple microbial data from a biological sample collected from an object; A step of selecting microbial-related features to be used in a machine learning model from among a plurality of extracted microbial data based on a preset feature selection algorithm; A step of training a machine learning model using selected microbial-related features; and A step of diagnosing and classifying migraine by inputting multiple microbial data extracted from microorganisms collected from the subject of examination into the learned machine learning model, The above microbial related features are: The total content and relative abundance of at least one microorganism selected from a first group of microorganisms consisting of microorganisms of the genus Gemella, microorganisms of the genus Streptococcus, microorganisms of the genus Granulicatella, microorganisms of the genus Rothia, microorganisms of the genus Mogibacterium, and microorganisms of the genus Schaalia; and A method comprising the total content and relative abundance of at least one microorganism selected from a second group of microorganisms consisting of microorganisms of the genus Alloprevotella, microorganisms of the genus Veillonella, microorganisms of the genus Haemophilus, microorganisms of the genus Selenomonas, microorganisms of the genus Campylobacter, microorganisms of the genus Cardiobacterium, microorganisms of the genus Megasphaera, and microorganisms of the genus Kingella. A method according to claim 17, wherein the machine learning model is selected from among a linear regression analysis (LRA) model, a random forest model, a generalized linear (GLMNET) model, a gradient boosting model, and an extreme gradient boost (XGB) model. In claim 17, A method for diagnosing migraine when, in a plurality of microbial data extracted from microorganisms collected from the subject of the test, the relative abundance of at least one microorganism selected from the first microbial group increases and / or the relative abundance of at least one microorganism selected from the second microbial group decreases.

Images