A method for constructing a blood glucose improvement prediction model of an individual after a dietary fiber intervention and application

Abstract

The application discloses a kind of individual blood glucose improvement prediction model after dietary fiber intervention The method for constructing is characterized in that, include the following steps: S1: collect the metabolite before intervention of subject, extract DNA, and obtain the original abundance matrix of Amplicon Sequence Variants (ASVs) by analysis;S2: construct LightGBM (Light Gradient-Boosting Machine) classification model;S3: adopt stratified 10-fold cross-validation method to create training set and test set;S4: construct microorganism group-fiber score formula, blood glucose benefit score;S5: according to the corresponding relationship of microorganism group-fiber score and blood glucose benefit score, predict by microorganism group-fiber score.The model of the application can predict whether the subject will benefit from dietary fiber intervention, help the subject to select effective treatment strategy before intervention begins.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of bioengineering and intestinal health, and particularly relates to the study of intestinal flora response to dietary fiber intervention, the evaluation and prediction of individual intervention results, and in particular to a method for constructing an individual blood glucose improvement prediction model after dietary fiber intervention and application. BACKGROUND

[0002] The incidence of diabetes is increasing year by year, and has become a major threat to the health of the nation. Intestinal flora and dietary fiber play an important role in the intervention and treatment of diabetic patients, not only in the intake of dietary fiber to regulate blood glucose levels by slowing down the digestion and absorption of carbohydrates, thereby helping to control blood glucose levels, but also in the production of short-chain fatty acids (SCFA). Intestinal flora produces acetic acid and butyric acid by fermenting dietary fiber, and these short-chain fatty acids can affect the release of intestinal hormones (such as GLP-1 and PYY), thereby regulating blood glucose levels. Certain dietary fiber components can selectively promote the growth of beneficial flora, such as short-chain fatty acid-producing flora, thereby improving the metabolic health of the host. Therefore, intestinal flora can affect the glucose metabolism of the host through multiple pathways, and an imbalance of intestinal flora is closely related to the occurrence and development of diabetes.

[0003] Dietary fiber intervention is an intervention method targeting intestinal flora, which can improve human health by regulating intestinal flora. Certain members of intestinal flora can ferment and utilize dietary fiber to produce products beneficial to human health. More and more evidence shows that the effect of dietary intervention on health shows great individual differences, which is related to the individualized response of intestinal flora to diet. Since humans do not have the ability to digest and utilize dietary fiber, the efficacy of dietary fiber intervention depends more on the individualization of intestinal flora.

[0004] Many dietary fiber intervention studies have observed that different individuals benefit differently after receiving dietary fiber intervention, and not all subjects benefit from dietary fiber intervention. Specifically, the current dietary fiber program cannot be targeted, and the dietary program prescribed by doctors may be the same, but due to individual differences, the actual effect is different, so for the same dietary program, some patients have good improvement effect, and some patients have poor improvement effect.

[0005] At present, there is no successful model based on intestinal flora to predict the benefit of individuals. If an individual blood glucose improvement prediction model after dietary fiber intervention can be achieved, it will be very meaningful to accurately evaluate the effectiveness of individuals in advance to some extent. SUMMARY

[0006] In view of the above-mentioned deficiencies in the prior art, the present invention provides a method for constructing and applying an individual blood glucose improvement prediction model after dietary fiber intervention. The gut microbiota before dietary fiber intervention successfully predicted the improvement of blood glucose in different individuals after receiving dietary fiber intervention.

[0007] The specific technical solution adopted in this invention is as follows:

[0008] In a first aspect of the present invention, a method for constructing an individual blood glucose improvement prediction model after dietary fiber intervention is provided, comprising the following steps:

[0009] S1: Collect metabolites from subjects before intervention, extract DNA, and analyze to obtain the raw abundance matrix of amplicon sequence variants (ASVs);

[0010] S2: Construct a LightGBM (Light Gradient-Boosting Machine) classification model;

[0011] S3: Use a hierarchical 10-fold cross-validation method to create training and test sets;

[0012] S4: Constructing a microbiome-fiber scoring formula and a glycemic benefit score;

[0013] S5: Based on the correspondence between the microbiome-fiber score and the glycemic benefit score, prediction is made using the microbiome-fiber score.

[0014] In some specific implementations, in S1, after DNA extraction, high-throughput amplicon sequencing of the V3-V4 region of the 16S rRNA gene is performed using a sequencing platform, and data analysis is performed using software to obtain the original abundance matrix of amplicon sequence variants.

[0015] In some specific implementations, in S2, a LightGBM (Light Gradient-Boosting Machine) classification model is constructed based on the abundance of amplicon sequence variants with a common rate of more than 20% in the subjects.

[0016] In some specific implementations, in S3, the dataset is initially divided into 10 mutually exclusive subsets of equal size. In each round, one subset is used as the test set, and the remaining subsets are used as the training set. The final result is obtained by averaging the results of all 10 subsets. The area under the receiver operating characteristic curve (ROC-AUC) is used as a metric to quantify the classifier's performance.

[0017] In some specific implementations, for feature selection, the average feature importance of hierarchical 10-fold cross-validation is first calculated, followed by a weighted average feature importance calculation, where the weighting coefficients are the frequencies of feature importance greater than 0. Then, based on the stepwise forward selection method of hierarchical 10-fold cross-validation, the optimal number of features is determined by ranking them according to their weighted average importance.

[0018] After determining the optimal number of features, a LightGBM classification model is constructed using these features. SHAP (Shapley additive explanations) is then used to calculate the contribution of each feature to the prediction result for each sample in the LightGBM classification model (i.e., the SHAP value). SHAP is a method used to interpret the prediction results of machine learning models; by calculating the contribution of each feature to the model's prediction results, it provides global and local explanations for the model.

[0019] The SHAP value is calculated based on the following formula

[0020]

[0021] Φi represents the SHAP value of feature i. F represents the set of all features. S represents the subset of features that does not include feature i. XS represents the values ​​of the input features in the subset S. fS(XS) represents the predicted output of the model given the subset S.

[0022] In this study, the shapviz function in the R package shapviz (version 0.9.5) was used to calculate the SHAP value of each feature in the LightGBM classification model.

[0023] In some specific implementation schemes, based on the SHAP values ​​of ASVs calculated above, the microbiome-fiber score (MFS) formula is constructed as follows:

[0024]

[0025] Among them, MFS i S represents the microbiome-fiber score of individual i. ij S is the MFS value of the j-th ASV in individual i, and n is the total number of ASVs; ij Defined as: if the contribution value (i.e., SHAP value) of the j-th ASV in individual i is greater than 0, then S ij =1; otherwise, S ij =0.

[0026] Furthermore, the contribution value of the j-th ASV in individual i is obtained using the SHAP (Shapley additive explanations) method. For each sample and each feature, the model generates a SHAP value; the contribution value is the SHAP value.

[0027] In some specific implementations, the blood glucose benefit score is obtained by adding up the scores of blood glucose indicators, including fasting plasma glucose (FPG), two-hour postprandial plasma glucose (PBG), and glycated hemoglobin (HbA1c).

[0028] Furthermore, the blood glucose index score is calculated as follows: if the change in blood glucose index (%change) is less than zero, 1 point is awarded; otherwise, 0 points are awarded. The change in blood glucose index (%change) = (post-intervention index - baseline index) ÷ baseline index × 100%.

[0029] In some specific implementation schemes, individuals with a blood glucose benefit score greater than or equal to 2 are considered high responders, while those with a score less than 2 are considered low responders.

[0030] In some specific implementations, MFS is divided into three ranges; the first range includes subjects with an MFS value of less than or equal to 17, who only have a glycemic benefit score of 0 and 1, indicating that these subjects have limited or no glycemic benefit after receiving dietary fiber intervention; the second range includes subjects with an MFS value of greater than or equal to 18 and less than or equal to 22, covering all glycemic benefit scores, indicating that these subjects have an unclear glycemic benefit after receiving dietary fiber intervention; the third range includes subjects with an MFS value of greater than or equal to 23, who have a glycemic benefit score of 2 and 3, indicating that these subjects can benefit from dietary fiber intervention.

[0031] In a second aspect of the invention, an individual blood glucose improvement prediction model is used to predict whether a subject will benefit from dietary fiber intervention.

[0032] The application of individual glycemic improvement prediction models after dietary fiber intervention includes the following steps:

[0033] S1: Collect fecal samples from subjects before intervention, extract DNA, and use a sequencing platform to perform high-throughput amplicon sequencing on the V3-V4 region of the 16S rRNA gene. Use software to perform data analysis to obtain the raw abundance matrix of amplicon sequence variants (ASVs).

[0034] S2: The LightGBM (Light Gradient-Boosting Machine) classification model described in claim 3 is adopted;

[0035] S3: Construct a microbiome-fiber scoring formula according to the construction method of claim 4, and make predictions based on the scores.

[0036] In a further application, in step S1, after DNA extraction, high-throughput amplicon sequencing of the V3-V4 region of the 16S rRNA gene is performed using a sequencing platform, and data analysis is performed using software to obtain the original abundance matrix of amplicon sequence variants.

[0037] In some specific implementations, the number of amplicon sequence variants is 44.

[0038] Compared with the prior art, the present invention has the following beneficial effects:

[0039] A machine learning model based on baseline gut microbiota characteristics successfully predicted whether subjects would benefit from dietary fiber intervention, and this prediction was validated in an external validation cohort. This demonstrates the model's strong applicability across diverse populations. This gut microbiota-based model provides a non-invasive way for subjects to select effective treatment strategies before intervention begins.

[0040] The following will further explain the concept and technical effects of the present invention in conjunction with the accompanying drawings, so as to fully understand the purpose, features and effects of the present invention. Attached Figure Description

[0041] Figure 1 This is a flowchart of a population cohort intervention process. Among them, Figure 1 'a' is the flowchart for discovering the intervention queue; Figure 1 b is the flowchart for the verification queue intervention.

[0042] Figure 2 This refers to the classification of blood glucose improvement among subjects in a population cohort. Among them, Figure 2 a represents the classification of blood glucose improvement among the subjects in the discovery cohort; Figure 2 b represents the classification of blood glucose improvement among subjects in the validation cohort.

[0043] Figure 3 It is an ASV diagram that can predict whether a participant is a high or low responder. Among them, Figure 3 a is the average ROC-AUC plot for different numbers of ASVs. Figure 3 b is a weighted average importance and abundance plot of ASVs selected by the LightGBM model in the discovery cohort; the clustering tree shows the associations between ASVs, which are determined by Spearman correlation coefficient based on the average abundance of all subjects; the heatmap shows the average abundance of each ASV in low-responders and high-responders (z-score transformation). Figure 3c is the abundance plot of ASVs selected by the LightGBM model in the validation cohort; the clustering tree shows the associations between ASVs, which are determined by Spearman correlation coefficient based on the mean abundance of all subjects; the heatmap shows the mean abundance of each ASV in low-responders and high-responders (z-score transformation).

[0044] Figure 4 This is a graph showing the model prediction results for the discovery queue and the validation queue. Figure 4 a) is a receiver operating characteristic (ROC) curve of the discovery cohort based on 10-fold cross-validation; the mean of the area under the receiver operating characteristic curve (AUC) is shown as mean ± standard deviation (SD). Figure 4 b is an ordered logistic regression plot showing the correlation between MFS and glycemic benefit score in the cohort; the bar chart (mean ± standard error (SEM)) shows the difference in MFS between low-responders and high-responders in the GPD study. Figure 4 c is the Receiver Operating Characteristic (ROC) curve of the validation cohort. Figure 4 d is an ordered logistic regression plot showing the correlation between MFS and glycemic benefit score in the validation cohort; the bar chart shows the difference in MFS between low-responders and high-responders in the GLC study.

[0045] Figure 5 This is a graph showing the correspondence between MFS and glycemic benefit score. Detailed Implementation

[0046] To make the technical means, inventive features, objectives, and effects of the invention readily understandable, the invention is further described below with reference to specific illustrations. However, the invention is not limited to the embodiments described below.

[0047] It should be noted that the scales, sizes, etc., shown in the accompanying drawings are only for the purpose of assisting those skilled in the art in understanding and reading the content disclosed in the specification, and are not intended to limit the conditions under which the present invention can be implemented. Therefore, they have no substantial technical significance. Any improvements that do not involve creative effort, without affecting the effects and objectives that the present invention can produce, should still fall within the scope of the technical content disclosed in the present invention.

[0048] Example 1 Model Construction

[0049] (I) Discovery Cohort Recruitment and Intervention Experiment

[0050] The flowchart for queue intervention is as follows: Figure 1As shown in figure a. We recruited 802 volunteers with prediabetes (as defined by the World Health Organization) in East China and randomly assigned them to two groups: a control group (Group U, n = 393) and a dietary fiber supplement intervention group (Group W, n = 409). The control group maintained normal lifestyles; the dietary fiber supplement intervention group received 45 grams of dietary fiber supplements daily for 6 months, and ultimately 331 volunteers in Group W completed the 6-month intervention. A predictive model constructed using baseline gut microbiota abundance data from the cohort predicted the improvement in blood glucose levels after intervention in different individuals. The dietary fiber supplements mentioned above were from Shanghai Jiuben Technology Co., Ltd.

[0051] (II) Classification of Blood Glucose Improvement

[0052] Based on the change in blood glucose levels (%change) after intervention, participants in the discovery cohort were divided into low-responders and high-responders. Blood glucose levels included fasting plasma glucose (FPG), 2-hour postprandial blood glucose (PBG), and glycated hemoglobin (HbA1c). A change in any blood glucose level (%change) less than zero was scored as 1 point; otherwise, 0 points were scored. The scores for the three blood glucose levels were then added together to obtain the individual's blood glucose benefit score. Individuals with a blood glucose benefit score greater than or equal to 2 were considered high-responders, and those less than 2 were considered low-responders. The formula for calculating %change is as follows: %change = (Post-intervention level - Baseline level) ÷ Baseline level × 100%. The responder division in the discovery cohort is as follows: Figure 2 As shown in a.

[0053] (III) Constructing a Prediction Model

[0054] Fecal samples were collected from volunteers at baseline. DNA was extracted, and high-throughput amplicon sequencing of the V3-V4 region of the 16S rRNA gene in the feces was performed using Illumina's Miseq sequencing platform. Data analysis was conducted using Qiime 2 software. First, adapters and primers were removed from the sequences using the "Cutadapt" plugin. Sequences were then pruned using DADA2 based on sequencing quality. After filtering, noise reduction, chimera removal, and merging, the raw abundance matrix of amplicon sequence variants (ASVs) was obtained. Due to differences in sequencing depth among different samples, the ASV abundance table was transformed using relative log expression (RLE), and the transformed abundance information was used for subsequent analysis.

[0055] Based on the abundance of ASVs with a common rate exceeding 20% ​​in the W group of the discovery cohort, a LightGBM (Light Gradient-Boosting Machine) classification model was constructed to distinguish between low-responders and high-responders. Hierarchical 10-fold cross-validation was used to create the training and test sets. The dataset was initially divided into 10 mutually exclusive subsets of equal size. In each round, one subset was used as the test set, and the remaining subsets were used as the training set. The final result was obtained by averaging the results of all 10 subsets. The Area Under the Receiver Operating Characteristic (ROC-AUC) was used as a metric to quantify the classifier's performance. The average ROC-AUC of the hierarchical 10-fold cross-validation was considered a measure of the model's accuracy.

[0056] For feature selection, the average feature importance of hierarchical 10-fold cross-validation is first calculated, followed by a weighted average feature importance calculation, where the weighting coefficients are the frequencies of feature importance greater than 0. Then, based on the stepwise forward selection method of hierarchical 10-fold cross-validation, the optimal number of features is determined by ranking them according to their weighted average importance. The ROC-AUC for different numbers of ASVs and the number, abundance, and taxonomic position of the finally selected ASVs are shown below. Figure 3 As shown in a, 3b.

[0057] The hyperparameters of the classification model were optimized through hierarchical 10-fold cross-validation grid search and manual tuning. The hyperparameters "num_leaves", "learning_rate", "max_depth", and "n_estimators" represent the number of leaves in the decision tree, the iteration rate, the maximum depth of the decision tree, and the number of iterations to improve, respectively. Following the recommendations in the LightGBM official documentation (https: / / lightgbm.readthedocs.io / en / latest / index.html), the hyperparameter ranges were set as follows: learning_rate [0.01, 0.1], n_estimators [100, 1000], max_depth [3, 7], num_leaves [7, 127]. The final model parameters were: learning_rate = 0.05, n_estimators = 115, max_depth = 7, num_leaves = 80.

[0058] The LightGBM classification model was built using the lightgbm package (version 3.3.5). Hierarchical 10-fold cross-validation was performed using the createMultiFolds function of the caret package (version 6.0-94). Feature importance was obtained using the lgb.importance function of the lightgbm package. Hyperparameter tuning used the tidymodels package (version 1.2.0) and the bonsai package (version 0.2.1). The pROC package (version 1.18.5) was used for ROC curve analysis, and the DeLong method was used to calculate the area under the curve.

[0059] Based on the above method, we identified 44 ASVs ( Figure 3 b. The DNA sequence is shown in Table 1. These 44 ASVs can be used to effectively distinguish between low-responders and high-responders.

[0060] Table 1. Sequence listing of 44 amplicon sequence variants

[0061]

[0062]

[0063]

[0064]

[0065]

[0066]

[0067]

[0068]

[0069]

[0070]

[0071]

[0072]

[0073]

[0074]

[0075]

[0076]

[0077]

[0078]

[0079] To analyze the contribution of ASVs to the classification model, we used the SHAP (Shapley additive explanations) method to visualize their contributions. SHAP is a game-theoretic approach used to analyze the contribution of features to a classification model. For each predicted sample and each feature, the model generates a SHAP value, representing the contribution of each feature. We used the shapviz function from the shapviz package (version 0.9.2) to calculate the SHAP value for each ASV in each sample. To quantify an individual's response to dietary fiber, we constructed a microbiome-fiber score (MFS) based on the SHAP values ​​of these 44 ASVs in each subject. The formula for constructing the microbiome-fiber score (MFS) is shown below:

[0080]

[0081] Among them, MFS i S represents the microbiome-fiber score of individual i. ij S is the MFS value of the j-th ASV in individual i, and n is the total number of ASVs. ij Defined as: if the SHAP value of the j-th ASV in individual i is greater than 0, then S ij =1; otherwise, S ij =0.

[0082] The results showed that MFS was significantly higher in high-responders than in low-responders, and MFS was significantly positively correlated with glycemic benefit score. Figure 4 (b) This demonstrates that MFS based on gut microbiome characteristics (44 ASVs) can be used to predict and quantify individual responses to dietary fiber interventions in prediabetic subjects in the discovery cohort. Regression analysis between MFS and glycemic benefit scores was performed using the lrm function in the rms package (version 6.8-0).

[0083] We further divided MFS into three ranges based on the correlation between MFS and glycemic benefit scores. The first range includes MFS values ​​less than or equal to 17, encompassing only subjects with glycemic benefit scores of 0 and 1, indicating that these volunteers experienced limited or no glycemic benefit after dietary fiber intervention. Figure 5The second range, including subjects with MFS values ​​greater than or equal to 18 and less than or equal to 22, encompassed all glycemic benefit scores, indicating that the glycemic benefit of these volunteers after receiving dietary fiber intervention was unclear. Figure 5 The third range, including subjects with an MFS value greater than or equal to 23, encompassed subjects with a glycemic benefit score of 2 and 3, indicating that these volunteers could benefit from the dietary fiber intervention. Figure 5 ).

[0084] Example 2 Model Application

[0085] (I) Recruitment and Intervention Experiments of the Verification Cohort

[0086] The flowchart for verifying queue intervention is as follows: Figure 1 As shown in b. The predictive model constructed for the discovery cohort was validated in an independent population cohort. Twenty volunteers were recruited in Shanghai, China, including 10 overweight or obese participants (3 of whom had prediabetes) and 10 participants with type 2 diabetes. All volunteers received a dietary fiber supplementation intervention. The intervention lasted for 14 days, with each participant consuming 18 grams of dietary fiber daily for the first 7 days and 36 grams daily for the following 7 days. Ultimately, 19 volunteers completed the intervention. Furthermore, volunteers (n=2) who could not be accurately grouped due to missing blood glucose data were excluded. The same dietary fiber supplement was used in the validation cohort as in the discovery cohort.

[0087] (II) Classification of Blood Glucose Improvement

[0088] Volunteers were categorized into low-responders and high-responders based on their percentage change (%change) of blood glucose indicators after intervention. Blood glucose indicators included fasting plasma glucose (FPG), mean continuous glucose (GCM), and glycated hemoglobin (HbA1c). A change (%change) less than zero was scored as 1 point, otherwise 0 points. The scores for the three indicators were summed to obtain the individual's glycemic benefit score. Individuals with a glycemic benefit score greater than or equal to 2 were considered high-responders, and those less than 2 were considered low-responders. In the validation cohort, continuous glucose monitoring (CGM) was used instead of postprandial two-hour blood glucose (PBG) since it was not measured. The formula for calculating %change is: %change = (post-intervention indicator - baseline indicator) ÷ baseline indicator × 100%. The responder classification in the validation cohort is as follows: Figure 2 As shown in b.

[0089] (III) Model Prediction Operation

[0090] The steps for model prediction are as follows:

[0091] 1. DNA was extracted from fecal samples of the validation cohort using the same method as the discovery cohort. High-throughput amplicon sequencing of the V3-V4 region of the 16S rRNA gene in feces was performed using Illumina’s Miseq sequencing platform.

[0092] 2. Sequencing data were processed using the same method as the discovery cohort to obtain the abundance matrix of amplicon sequence variants (ASVs).

[0093] 3. Search the validation queue for the 44 ASVs found in the discovery queue, and use the abundance information of these 44 ASVs in the validation queue for subsequent analysis.

[0094] 4. The constructed LightGBM (LightGradient-Boosting Machine) classification model was trained using the abundance information of the 44 ASVs in the discovery queue. This was achieved using the lgb.train function of the lightgbm package (version 3.3.5).

[0095] 5. Input the abundance information of the 44 ASVs in the validation queue into the trained model, and use the abundance information of the 44 ASVs in the validation queue to predict the low-responders and high-responders in the validation queue. This is achieved using the predict function of the lightgbm package (version 3.3.5).

[0096] 6. Input the grouping information of the validation queue predicted by the model and the actual grouping information of the validation queue into the roc function of the pROC package (version 1.18.5) to calculate the ROC curve and the area under the ROC curve.

[0097] 7. In the validation queue, the shapviz function from the shapviz package (version 0.9.5) is used to calculate the SHAP value of each ASV in each sample.

[0098] 8. Using the same method as the discovery cohort, calculate the microbiome-fiber score (MFS) for each sample in the validation cohort using the SHAP value obtained in the previous step.

[0099] 9. Use the lrm function in the rms package (version 6.8-0) to perform regression analysis between MFS and glycemic benefit score.

[0100] (III) Model Prediction Results and Validation Queue Results

[0101] The effectiveness of the classification model was verified using data from the validation queue. It was found that the constructed LightGBM classification model could also effectively distinguish between high and low responders in the validation queue. Figure 4 c). Using the same method to construct the microbiome-fiber score (MFS), it was also found in the validation cohort that high-responder MFS was significantly higher than low-responder MFS, and MFS was significantly positively correlated with glycemic benefit score in the validation cohort. Figure 4 d). Figure 5 This is a graph showing the correspondence between MFS and glycemic benefit scores. MFS is divided into three score ranges, reflecting the degree of benefit from dietary fiber intervention, which is consistent with the glycemic benefit score.

[0102] It is evident that this model can be used to predict the benefits that different individuals will receive after dietary fiber intervention.

[0103] The preferred embodiments of the present invention have been described in detail above. It should be understood that those skilled in the art can make numerous modifications and variations based on the concept of the present invention without creative effort. Therefore, all technical solutions that can be obtained by those skilled in the art based on the concept of the present invention through logical analysis, reasoning, or limited experimentation on the basis of existing technology should be within the scope of protection defined by the claims.

Claims

1. A method for constructing a blood glucose improvement prediction model of an individual after a dietary fiber intervention, characterized by, Comprising the following steps: S1: Collecting metabolites before intervention of subjects, extracting DNA, and analyzing to obtain an original abundance matrix of Amplicon Sequence Variants (ASVs); S2: Based on the abundance of the Amplicon Sequence Variants with a common rate of more than 20% in the subjects, a LightGBM (Light Gradient-Boosting Machine) classification model is constructed; S3: A stratified 10-fold cross-validation method is used to create a training set and a test set; S4: A microbiome-fiber score formula and a blood glucose benefit score are constructed, wherein the microbiome-fiber score (MFS) formula is constructed according to the Amplicon Sequence Variants identified in the S3 step as follows: where MFS i represents the microbiome-fiber score value for individual i, S ij is the MFS value for the jth ASV in individual i, and n is the total number of ASVs; S ij is defined as: S ij = 1 if the contribution value of the jth ASV in individual i is greater than 0; otherwise, S ij = 0. And the blood glucose benefit score is obtained by adding the scores of blood glucose indicators, including fasting plasma glucose (FPG), postprandial blood glucose (PBG), and glycosylated hemoglobin (HbA1c); If the change value (%change) of the blood glucose indicator is less than zero, it is scored as 1, otherwise it is scored as 0; the change value (%change) of the blood glucose indicator = (post-intervention indicator - baseline indicator) ÷ baseline indicator × 100%; S5: According to the corresponding relationship between the microbiome-fiber score and the blood glucose benefit score, the microbiome-fiber score is used for prediction.

2. The method of claim 1, wherein the method is for constructing a model for predicting improvement in blood glucose in an individual after dietary fiber intervention. In the S1 step, after extracting DNA, high-throughput amplicon sequencing of the V3-V4 segment of the 16S rRNA gene is performed using a sequencing platform, and data analysis is performed using software to obtain an original abundance matrix of Amplicon Sequence Variants.

3. The method of claim 1, wherein the method is for constructing a model for predicting improvement in blood glucose in an individual after dietary fiber intervention. The contribution value of the jth ASV in individual i is obtained using the SHAP (Shapley additive explanations) method, which generates a SHAP value for each sample and each feature; the contribution value is the SHAP value.

4. The method of claim 1, wherein the method is for constructing a model for predicting improvement of blood glucose in an individual after dietary fiber intervention. Individuals with a blood glucose benefit score greater than or equal to 2 are high responders, and individuals with a score less than 2 are low responders.

5. The method of claim 1, wherein the method is for constructing a model for predicting improvement of blood glucose in an individual after dietary fiber intervention. The MFS is divided into three ranges; the first range includes MFS values less than or equal to 17, and only includes subjects with blood glucose benefit scores of 0 and 1, indicating that these subjects have limited or no blood glucose benefit after receiving dietary fiber intervention; the second range includes subjects with MFS values greater than or equal to 18 and less than or equal to 22, involving all blood glucose benefit scores, indicating that the blood glucose benefit of these subjects is unclear after receiving dietary fiber intervention; the third range includes subjects with MFS values greater than or equal to 23, including subjects with blood glucose benefit scores of 2 and 3, indicating that these subjects can benefit from dietary fiber intervention.

6. Use of a blood glucose improvement prediction model for an individual after a dietary fiber intervention, characterized in that, The individual blood glucose improvement prediction model after the dietary fiber intervention is obtained by the construction method of any one of claims 1-5, and the application of the model includes predicting whether the subjects will benefit from dietary fiber intervention.

7. Use according to claim 6, characterized in that, Comprising the following steps: S1: Collecting fecal samples of subjects before intervention, extracting DNA, using a sequencing platform to perform high-throughput amplicon sequencing on the V3-V4 segment of the 16S rRNA gene, and using software to analyze the data to obtain the original abundance matrix of amplicon sequence variants (ASVs); S2: using the LightGBM (Light Gradient-Boosting Machine) classification model of claim 3; S3: constructing the microbiome-fiber scoring formula according to the construction method of claim 4, and making a prediction according to the score.

8. Use according to claim 7, characterized in that, In the S1 step, after extracting DNA, the V3-V4 segment of the 16S rRNA gene is sequenced using a sequencing platform for high-throughput amplicon sequencing, and software is used for data analysis to obtain the original abundance matrix of amplicon sequence variants.

9. Use according to claim 7, characterized in that, The amplicon sequence variants are 44.

Images