Acute pancreatitis intestinal micro-ecological marker and application thereof
By using 18 gut microbiota biomarkers and a random forest model to construct a prediction model, the problem of accuracy in the early diagnosis of acute pancreatitis was solved, achieving high accuracy and high sensitivity in the prediction of acute pancreatitis.
Patent Information
- Application Number
- CN202511086460.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-08-04
- Publication Date
- 2025-11-21
AI Technical Summary
Current technologies lack biomarkers that can accurately predict the severity and prognosis of acute pancreatitis in its early stages, resulting in some high-risk patients not receiving timely intervention and affecting treatment outcomes.
Using 18 gut microbial biomarkers, including Bacteroides, Clostridium, Trichophyton, and Agathobacter, a predictive model was constructed using a random forest model. Combined with 16S rRNA gene sequencing data, an acute pancreatitis diagnostic system and kit were developed for the early diagnosis of acute pancreatitis.
The model achieved high accuracy and sensitivity in predicting acute pancreatitis, with an AUC of 0.98, precision of 0.955, and recall of 0.877, enabling early diagnosis of acute pancreatitis.
Smart Images
Figure CN120989229A_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of disease diagnosis technology, specifically relating to an intestinal microecological marker for acute pancreatitis and its application. Background Technology
[0002] Acute pancreatitis (AP) is a common digestive emergency with a high mortality rate. The main causes of death include infected pancreatic necrosis and multiple organ dysfunction syndrome. Currently, there is a lack of biomarkers that can accurately predict the severity and prognosis of AP in its early stages, leading to some high-risk patients not receiving timely intervention and severely impacting treatment outcomes.
[0003] Recent studies have found that gut microbiota dysbiosis is closely related to the occurrence and development of acute pancreatitis (AP). Early-stage AP patients show a decrease in gut microbiota diversity and abundance, but no clinically applicable methods or criteria for detecting and diagnosing AP gut microbiota biomarkers have yet been established. Summary of the Invention
[0004] In view of this, the purpose of the present invention is to provide a gut microbiota biomarker for acute pancreatitis, which has the characteristic of predictive accuracy and can be used for the early diagnosis of acute pancreatitis.
[0005] This invention provides a gut microbiota biomarker for acute pancreatitis, including the following biomarkers:
[0006] Bacteroides, Clostridium sensu strain 1, Lachnospira, Agathobacter, Roseburia, Prevotella, Escherichia-Shigella, Faecalibacterium, Prevotella 9, Veillonella, Ruminococcus 2, Anaerococcus, Romboutsia, Blautia, Enterococcus, Finegoldia, Eubacterium eligensgroup, and Clostridium sensu stricto1.
[0007] In this invention, the intestinal microecological marker for acute pancreatitis preferably includes a first combination of the marker and Peptoniphilus or a second combination of the marker and Peptoniphilus and Acinetobacter.
[0008] This invention provides the application of the aforementioned intestinal microecological biomarkers for acute pancreatitis in constructing a predictive model for acute pancreatitis.
[0009] Preferably, the method for constructing the acute pancreatitis prediction model includes using a random forest model.
[0010] This invention provides a diagnostic system for acute pancreatitis, comprising the following connected modules:
[0011] The data acquisition module is used to collect sequencing data of the intestinal microecological markers of acute pancreatitis in the sample to be tested;
[0012] The data analysis module is used to analyze and predict the sequencing data of the data acquisition module using the acute pancreatitis prediction model constructed in the application, and obtain the prediction results.
[0013] The diagnostic output module is used to output the prediction results obtained by the data analysis module to the terminal for display.
[0014] Preferably, the sequencing data of the intestinal microecological markers for acute pancreatitis includes the 16S rRNA gene sequencing data of the intestinal microecological markers for acute pancreatitis.
[0015] Preferably, the analysis and prediction method of the acute pancreatitis prediction model is to determine the risk of disease when the predicted value is ≥0.5, and to determine no risk of disease when the predicted value is <0.5.
[0016] The present invention provides an acute pancreatitis diagnostic device, including an acute pancreatitis diagnostic system equipped with the above-described technical solution.
[0017] This invention provides the application of a reagent for detecting gut microbiota markers of acute pancreatitis in the preparation of a kit for predicting acute pancreatitis.
[0018] Preferably, the reagent includes primers for amplifying the V3-V4 region of the bacterial 16S rRNA gene;
[0019] The primers used for amplifying the V3-V4 region of the bacterial 16S rRNA gene include a forward primer with a nucleotide sequence as shown in SEQ ID NO:1 and a reverse primer with a nucleotide sequence as shown in SEQ ID NO:2.
[0020] This invention provides a gut microbiota biomarker for acute pancreatitis, comprising the following biomarkers: Bacteroides, Clostridium sensu strain 1, Lachnospira, Agathobacter, Roseburia, Prevotella, Escherichia-Shigella, Faecalibacterium, Prevotella 9, Veillonella, Ruminococcus 2, Anaerococcus, Romboutsia, Blautia, Enterococcus, Finegoldia, Eubacterium eligens group, and Clostridium sensu stricto1. This invention screened 18 significantly differentially expressed dominant gut microbiota from healthy and acute pancreatitis samples. A prediction model constructed using a random forest model was used to evaluate the accuracy of diagnosing acute pancreatitis based on these biomarkers. The results showed that the prediction model had an AUC of 0.98, precision of 0.955, and recall of 0.877. Therefore, these microbial biomarkers can accurately predict acute pancreatitis and can be used for early clinical diagnosis of acute pancreatitis.
[0021] The intestinal microecological biomarkers for acute pancreatitis provided by this invention further include *Peptoniphilus* or a combination of *Peptoniphilus* and *Acinetobacter*. Although *Peptoniphilus* and *Acinetobacter* have low weights in the constructed prediction model, the constructed prediction model has high diagnostic accuracy, with an AUC value of 0.98 for all models, and also exhibits high precision and recall. Attached Figure Description
[0022] Figure 1 The results show the differential expression of 20 microbial biomarkers screened.
[0023] Figure 2 The confusion matrix result for Model 1;
[0024] Figure 3 The classification report results are for the predictions of Model 1;
[0025] Figure 4 The ROC curve results for Model 1
[0026] Figure 5 The feature importance results predicted by Model 1;
[0027] Figure 6 The classification report and ROC curve results for Model 2's predictions;
[0028] Figure 7 The classification report and ROC curve results for Model 3 predictions. Detailed Implementation
[0029] This invention provides a gut microbiota biomarker for acute pancreatitis, including the following biomarkers:
[0030] Bacteroides, Clostridium sensu strain 1, Lachnospira, Agathobacter, Roseburia, Prevotella, Escherichia-Shigella, Faecalibacterium, Prevotella 9, Veillonella, Ruminococcus 2, Anaerococcus, Romboutsia, Blautia, Enterococcus, Finegoldia, Eubacterium eligensgroup, and Clostridium sensu stricto1.
[0031] In this invention, the markers are obtained through screening using the following method:
[0032] Using samples from patients diagnosed with acute pancreatitis and samples from healthy individuals who were not diagnosed with acute pancreatitis by doctors, the 16S rRNA gene was sequenced, yielding two sets of sequencing data.
[0033] The two sets of sequencing data were optimized for low-quality data to obtain the relative abundance data of bacterial genera in each sample. After differential analysis, two sets of significantly different bacterial genera were obtained, thus obtaining the intestinal microecological markers for acute pancreatitis.
[0034] In this invention, the sample is preferably feces. The primers used for 16S rRNA gene sequencing preferably include a forward primer with the nucleotide sequence shown in SEQ ID NO:1 and a reverse primer with the nucleotide sequence shown in SEQ ID NO:2. This invention does not impose any particular limitation on the sequencing method used; any sequencing platform well-known in the art can be employed.
[0035] In this invention, the difference analysis preferably includes the use of LEfSe, and the analysis method of LEfSe preferably includes linear discriminant analysis and effect size method.
[0036] In this invention, the following strains are included: Prevotella 9, Blautia, Faecalibacterium, Romboutsia, Agathobacter, Lachnospira, Ruminococcus 2, Eubacterium eligens group, Roseburia, Clostridium sensu strain 1, and Clostridium sensu stricto. 1. Expression was upregulated in patients with acute pancreatitis; while expression was downregulated in patients with acute pancreatitis of Bacteroides, Prevotella, Escherichia-Shigella, Veillonella, Anaerococcus, Enterococcus, and Finegoldia.
[0037] In this invention, the intestinal microecological biomarkers for acute pancreatitis preferably include a first combination of the biomarkers and *Peptoniphilus*, or a second combination of the biomarkers and *Peptoniphilus* and *Acinetobacter*. Both *Peptoniphilus* and *Acinetobacter* are downregulated in patients with acute pancreatitis. Although *Peptoniphilus* and *Acinetobacter* have relatively low weighting in the prediction model, constructing a prediction model using the aforementioned 18 biomarkers and 19 biomarkers formed by *Peptoniphilus*, or 20 biomarkers formed by combining *Peptoniphilus* and *Acinetobacter*, still achieves high prediction accuracy and improved sensitivity.
[0038] This invention provides the application of the aforementioned intestinal microecological biomarkers for acute pancreatitis in constructing a predictive model for acute pancreatitis.
[0039] In this invention, the method for constructing an acute pancreatitis prediction model preferably involves detecting the 16S rRNA of the acute pancreatitis gut microbiota biomarkers in training samples, analyzing the relative abundance data of the acute pancreatitis gut microbiota biomarkers, using the relative abundance data of the acute pancreatitis gut microbiota biomarkers of each sample as input data to construct a prediction model, and then using the relative abundance data of the acute pancreatitis gut microbiota biomarkers of sequencing samples to evaluate the accuracy of the prediction model.
[0040] In this invention, the method for constructing the acute pancreatitis prediction model preferably includes using a random forest model. When constructing the acute pancreatitis prediction model, the random forest model in the pycaret software is preferred; the specific parameters are n_estimators = 100, max_features = sqrt, max_depth = 30, and min_samples_leaf = 1.
[0041] This invention provides a diagnostic system for acute pancreatitis, comprising the following connected modules:
[0042] The data acquisition module is used to collect sequencing data of the intestinal microecological markers of acute pancreatitis in the sample to be tested;
[0043] The data analysis module is used to analyze and predict the sequencing data of the data acquisition module using the acute pancreatitis prediction model constructed in the application of the above technical solution, and obtain the prediction results.
[0044] The diagnostic output module is used to output the prediction results obtained by the data analysis module to the terminal for display.
[0045] In this invention, the sequencing data of the gut microbiota biomarker for acute pancreatitis preferably includes the 16S rRNA gene sequencing data of the gut microbiota biomarker for acute pancreatitis. The preferred method for analyzing and predicting the acute pancreatitis prediction model is to determine the risk of disease when the predicted value is ≥0.5, and to indicate no risk of disease when the predicted value is <0.5.
[0046] The present invention provides an acute pancreatitis diagnostic device, including an acute pancreatitis diagnostic system equipped with the above-described technical solution.
[0047] In this invention, the device further includes a sequencing device for the intestinal microecological markers of acute pancreatitis in the sample to be tested and / or a terminal device for storing and / or displaying the prediction results.
[0048] This invention provides the application of a reagent for detecting gut microbiota markers of acute pancreatitis in the preparation of a kit for predicting acute pancreatitis.
[0049] In this invention, the reagent preferably includes primers for amplifying the V3-V4 region of the bacterial 16S rRNA gene. The primers for amplifying the V3-V4 region of the bacterial 16S rRNA gene preferably include a forward primer with a nucleotide sequence as shown in SEQ ID NO:1 (TCGTCGGCAGCGTCAGATGTGTATAAGAGACAGCCTACGGGNGGC WGCAG) and a reverse primer with a nucleotide sequence as shown in SEQ ID NO:2 (GTCTCGTGGGCTCGGAGATGTGTATAAGAGACAGGACTACHVGGGTATC TAATCC).
[0050] The following detailed description, in conjunction with embodiments, illustrates an intestinal microecological marker for acute pancreatitis and its application provided by the present invention. However, these descriptions should not be construed as limiting the scope of protection of the present invention.
[0051] Example 1
[0052] A method for screening gut microbiota biomarkers in acute pancreatitis
[0053] 1. Sample Description
[0054] Fecal samples were collected from 240 hospital-diagnosed cases of acute pancreatitis and 241 healthy individuals. Samples from individuals who had taken antibiotics within the past 1-3 months were excluded. The acute pancreatitis samples were those diagnosed by a physician; the healthy group samples were from individuals examined and identified by a physician as not having acute pancreatitis and whose indicators were relatively normal. After collection, samples were preserved using dry ice for low-temperature transport and stored at -80°C.
[0055] 2.16S rRNA sequencing
[0056] Nucleic acid extraction was performed on samples from the acute pancreatitis group and the healthy group using magnetic beads (Surbiopure fecal nucleic acid extraction kit, Guangzhou Cybex Biotechnology Co., Ltd.). The extracted DNA was amplified in the V3-V4 region of the bacterial 16S rRNA gene using the TransStartFastPfu Fly DNAPolymerase kit (TransGenBiotech, Beijing) via the MiniAmp Plus Thermal Cycler (Thermo Fisher Scientific). The primer set consisted of 341F (5'-TCGTCGGCAGCGTCAGATGTGTATAAGAGACAGCCTACGGGNGGCW GCAG-3', SEQ ID NO:1) and 805R (5'-GTCTCGTGGGCTCGGAGATGTGTATAAGAGACAGGACTACHVGGGTA TCTAATCC-3', SEQ ID NO:2) with adapters. After library construction, sequencing was performed using an Illumina MiSeq instrument.
[0057] 3. Screening for dominant bacteria
[0058] The data was quality controlled using the dada2 tool in qiime2-2020.2. Some low-quality data were filtered out, and the remaining high-quality data were classified and annotated with reference to the Silva database to obtain the relative abundance data of the bacterial genus level of the samples. The LEfSe (linear discriminant analysis and effect size method) was used to analyze the differences between the acute pancreatitis group and the healthy group. Significantly different bacterial genera were identified as biomarkers. Twenty biomarkers were screened, specifically: *Bacteroides*, *Peptoniphilus*, *Clostridium sensu* strain 1, *Acinetobacter*, *Lachnospira*, *Agathobacter*, *Roseburia*, *Prevotella*, *Escherichia-Shigella*, *Faecalibacterium*, nine *Prevotella* strains, *Veillonella*, and *Ruminococcus*. 2) Anaerococcus, Romboutsia, Blautia, Enterococcus, Finegoldia, Eubacterium eligens group, Clostridium sensu stricto 1. The expression trends of these 20 markers are shown in [the original text]. Figure 1 .
[0059] Example 2
[0060] 1. Construction of an acute pancreatitis model
[0061] The samples diagnosed with acute pancreatitis by professional doctors were designated as the acute pancreatitis group, while the samples of people who were examined and identified by professional physicians as not having acute pancreatitis and whose various indicators were relatively normal were designated as the healthy group.
[0062] Both sets of samples were divided into training and test sets at a ratio of 7:3. The expression abundance of 20 biomarkers in the training samples was determined according to the method in Example 1, resulting in two sets of sequencing data.
[0063] Table 1. Sample size of the acute pancreatitis group and the healthy group.
[0064] Acute pancreatitis group Health Group Training samples 168 169 Test samples 73 72
[0065] Note: During the modeling process, the healthy group samples were labeled with 0, and the acute pancreatitis group samples were labeled with 1.
[0066] The random forest model in the pycaret software was used to train and model two sets of sequencing data, and the constructed random forest model was used to predict the test samples. The specific parameters used in the modeling were n_estimators=100, max_features=sqrt, max_depth=30, and min_samples_leaf=1.
[0067] See prediction results Figure 2 The confusion matrix results from the test samples show that, in the control group (72 test samples), 68 samples were correctly predicted and 4 samples were incorrectly predicted; in the acute pancreatitis group (73 test samples), 68 samples were correctly predicted and 5 samples were incorrectly predicted. Similarly, in the healthy group (72 test samples), 68 samples were correctly predicted; and in the acute pancreatitis group (73 test samples), 68 samples were correctly predicted.
[0068] The classification report results are predicted from the test samples. Figure 3 The results show that the precision, recall, and F1-score (F1) of the 72 control group samples were 93.2%, 94.4%, and 93.8%, respectively; while the precision, recall, and F1-score (F1) of the 73 acute pancreatitis samples were 94.4%, 93.2%, and 93.8%, respectively.
[0069] Recall, also known as sensitivity or true positive rate (TPR), refers to the proportion of actual diseased or positive samples that are correctly identified as positive by the test. It reflects the ability of a testing method to correctly identify the target object; high sensitivity means a low rate of false negatives.
[0070] Precision measures how many samples predicted as positive by a model are actually positive, reflecting the "purity" of the model's predictions. Precision values range from 0.0 to 1.0, where 1.0 indicates that all samples predicted as positive are actually positive (perfect precision), and 0.0 indicates that no samples predicted as positive are actually positive.
[0071] F1 is the harmonic mean of precision and recall, providing a single metric that balances the two. The calculation formula is shown in Formula I.
[0072]
[0073] Precision represents the accuracy rate, and recall represents the recall rate.
[0074] The classification report results are predicted from the test samples. Figure 4 As can be seen, both AUC and ROC reached 0.98, indicating a very good prediction effect.
[0075] The predicted feature importance results from the test samples (see) Figure 5 As can be seen, the top five genera are Romboutsia, Lachnospira, Blautia, Eubacterium eligens group, and Clostridium sensustricto 1.
[0076] In Example 3, the bacterial genus with the lowest weight (Acinetobacter) was removed from the feature importance ranking of Model 1, and the sequencing data of the remaining 19 biomarkers were remodeled according to the method of Model 1 described in Example 2 to obtain an acute pancreatitis prediction model, denoted as Model 2.
[0077] See results Figure 6 Compared to Model 1, the recall rate for the acute pancreatitis group (labeled 1) and the precision rate for the healthy group (labeled 0) showed a significant decrease, indicating that the accuracy of the acute pancreatitis prediction model constructed with 19 biomarkers was lower than that of Model 1 constructed with sequencing data from 20 biomarkers. However, the modeling results had an AUC value of 0.98, still demonstrating high diagnostic accuracy and applicability for acute pancreatitis prediction.
[0078] Example 4
[0079] From the feature importance ranking of Model 1, the two bacteria with the lowest weights (Acinetobacter and Peptoniphilus) are removed, and the model is rebuilt in the same way as Model 1 to obtain the acute pancreatitis prediction model, which is denoted as Model 3.
[0080] See results Figure 7 Compared to Model 1, the recall rate for the acute pancreatitis group (labeled 1) and the precision rate for the healthy group (labeled 0) showed a significant decrease, indicating that the accuracy of the acute pancreatitis prediction model constructed with 18 biomarkers was lower than that of Model 1 constructed with sequencing data from 20 biomarkers. However, the modeling results had an AUC value of 0.98, still demonstrating high diagnostic accuracy and applicability for acute pancreatitis prediction.
[0081] A comprehensive evaluation of the modeling results using different numbers of biomarkers showed that the modeling using 20 biomarkers yielded the best results, with precision, recall, and F1-Score (f1) all exceeding 93, and an AUC of 98%. Therefore, these 20 bacteria were identified as gut microbiota biomarkers for acute pancreatitis. Using these biomarkers, acute pancreatitis can be predicted relatively accurately. Specifically, they are: Bacteroides, Peptoniphilus, Clostridium sensu stricto 1, Acinetobacter, Lachnospira, Agathobacter, Roseburia, Prevotella, Escherichia-Shigella, Faecalibacterium, Prevotella 9, Veillonella, Ruminococcus 2, Anaerococcus, Romboutsia, Blautia, Enterococcus, Finegoldia, Eubacterium coprostanoligenes group, and Eubacterium eligens group. Using this method for the detection of acute pancreatitis is non-invasive and highly accurate, providing a new approach for the detection of acute pancreatitis.
[0082] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.
Claims
1. A gut microbiota biomarker for acute pancreatitis, characterized in that, Including the following markers: Bacteroides, Clostridium sensu strain 1, Lachnospira, Agathobacter, Roseburia, Prevotella, Escherichia-Shigella, Faecalibacterium, Prevotella 9, Veillonella, Ruminococcus 2, Anaerococcus, Romboutsia, Blautia, Enterococcus, Finegoldia, Eubacterium eligens group, and Clostridium sensu stricto1.
2. The intestinal microecological marker for acute pancreatitis according to claim 1, characterized in that, The intestinal microecological markers for acute pancreatitis include a first combination of the marker and *Peptoniphilus* or a second combination of the marker and *Peptoniphilus* and *Acinetobacter*.
3. The application of the intestinal microecological biomarker for acute pancreatitis as described in claim 1 or 2 in constructing a predictive model for acute pancreatitis.
4. The application according to claim 3, characterized in that, The method for constructing the acute pancreatitis prediction model includes using a random forest model.
5. A diagnostic system for acute pancreatitis, characterized in that, Includes the following connected modules: The data acquisition module is used to acquire sequencing data of the intestinal microecological markers of acute pancreatitis as described in claim 1 or 2 in the sample to be tested; The data analysis module is used to analyze and predict the sequencing data of the data acquisition module using the acute pancreatitis prediction model constructed in the application of claim 2 or 3, and obtain the prediction results; The diagnostic output module is used to output the prediction results obtained by the data analysis module to the terminal for display.
6. The acute pancreatitis diagnostic system according to claim 5, characterized in that, The sequencing data of the intestinal microbiota biomarkers for acute pancreatitis includes the 16S rRNA gene sequencing data of the intestinal microbiota biomarkers for acute pancreatitis.
7. The acute pancreatitis diagnostic system according to claim 5, characterized in that, The analysis and prediction method of the acute pancreatitis prediction model is to determine the risk of disease when the predicted value is ≥0.5, and not to determine the risk of disease when the predicted value is <0.
5.
8. A diagnostic device for acute pancreatitis, characterized in that, Includes the acute pancreatitis diagnostic system described in any one of claims 5 to 7.
9. The use of a reagent for detecting the intestinal microecological marker of acute pancreatitis as described in claim 1 or 2 in the preparation of a kit for predicting acute pancreatitis.
10. The application according to claim 9, characterized in that, The reagents include primers for amplifying the V3-V4 region of the bacterial 16S rRNA gene; The primers used for amplifying the V3-V4 region of the bacterial 16S rRNA gene include a forward primer with a nucleotide sequence as shown in SEQ ID NO:1 and a reverse primer with a nucleotide sequence as shown in SEQ ID NO:2.