Machine learning model for predicting traditional Chinese medicine dampness syndrome and application thereof

Abstract

The invention discloses a biomarker for diagnosing and / or predicting traditional Chinese medicine dampness syndromes. The biomarker comprises a body mass index, a leukocyte count, a neutrophil percentage, a mononuclear cell percentage, lactic dehydrogenase, high-density lipoprotein, apolipoprotein A, adiponectin, epsilon-lysine, pentosan, carboxyethyl lysine, formyl lysine, carboxymethyl lysine, hydroxyimidazolone and / or acetyl lysine. Based on the biomarker, a machine learning model for predicting the traditional Chinese medicine dampness syndrome is provided, when the machine learning model is used for predicting the traditional Chinese medicine dampness syndrome, diagnosis and prediction are carried out on the basis of the biomarker obtained through objective detection, standardized judgment of the traditional Chinese medicine dampness syndrome risk is achieved, the repeatability of diagnosis is remarkably improved, and the diagnosis efficiency is improved. And the subjective dependence barrier of traditional Chinese medicine dampness syndrome diagnosis is thoroughly broken through, and insufficient diagnosis consistency caused by dependence on doctor experience and limitation of fuzzy symptom expression, namely individual difference, is avoided.

Description

Technical Field

[0001] This application relates to the field of TCM diagnostic technology, and in particular to a machine learning model for predicting dampness syndrome in TCM and its application. Background Technology

[0002] Dampness syndrome refers to the syndrome caused by external dampness or abnormal metabolism of body fluids leading to dampness and turbidity obstruction. It is characterized by symptoms such as heaviness and soreness of the body, abdominal distension and diarrhea, slippery tongue coating and soft pulse. The core pathogenesis is that dampness and turbidity obstruct the flow of qi and clear yang.

[0003] Dampness syndrome is divided into external dampness and internal dampness: external dampness is mostly caused by a humid environment or wading through water in the rain, while internal dampness is caused by spleen deficiency and impaired digestion or improper diet; the two often have a causal relationship. Clinically, it is divided into superficial dampness syndrome, qi-level damp-heat syndrome, and spleen deficiency with dampness syndrome, etc. Treatment methods include aromatic dampness-resolving, heat-clearing and dampness-draining, and spleen-strengthening and dampness-draining. Dampness can also cause dampness-bi syndrome (stagnant arthralgia), which manifests as severe joint pain, and is divided into wind-dampness, cold-dampness, and damp-heat obstruction syndromes, corresponding to treatments such as dispelling wind and cold, clearing heat and eliminating dampness.

[0004] Currently, the clinical identification of dampness syndrome relies primarily on the subjective experience of traditional Chinese medicine (TCM) practitioners, which challenges the consistency and reproducibility of diagnoses. Although some auxiliary diagnostic scales for dampness syndrome in TCM have been developed to improve objectivity through standardized scoring of symptoms and signs, these scales are still indirect assessment tools based on patients' subjective feelings and physicians' experience, making it difficult to completely eliminate subjective factors. Furthermore, the sensitivity and specificity of these scales are often limited by the ambiguity of symptom descriptions and individual patient differences. More importantly, such scales struggle to reflect the internal biological changes of dampness syndrome in real time and quantitatively, and their revelation of microscopic pathological processes is insufficient.

[0005] Current research has found that abnormal lipid metabolism indicators (such as elevated total cholesterol and decreased APOA1) are associated with the severity of phlegm-dampness syndrome; immune function indicators such as monocyte subsets show changes in rheumatic diseases; and metabolic-related host-like structures have been identified as risk factors for metabolic diseases. However, most current studies are limited to examining the sporadic correlation between single or a few laboratory indicators (such as blood lipids and immune markers) and dampness syndrome, lacking a systematic integration of comprehensive indicators that reflect immune function, tissue metabolism, lipid metabolism, fluid balance and internal environment, as well as specific toxic substances.

[0006] Therefore, there is a need for a method that comprehensively considers multiple factors and indicators and can accurately predict dampness syndrome in traditional Chinese medicine, so as to achieve early screening, risk warning and efficacy evaluation of dampness syndrome, and provide scientific basis and technical support for the modernization and precision diagnosis and treatment of dampness syndrome in traditional Chinese medicine. Summary of the Invention

[0007] The purpose of this invention is to overcome the above-mentioned shortcomings of the prior art and provide a model for diagnosing and / or predicting dampness syndrome in traditional Chinese medicine and its application.

[0008] The first objective of this invention is to provide a set of biomarkers for use in the preparation of products for diagnosing and / or predicting dampness syndrome in traditional Chinese medicine.

[0009] A second objective of this invention is to provide a product for diagnosing and / or predicting dampness syndrome in Traditional Chinese Medicine.

[0010] A third objective of this invention is to provide a machine learning model for predicting dampness syndrome in Traditional Chinese Medicine.

[0011] The fourth objective of this invention is to provide a predictive system for predicting dampness syndrome in Traditional Chinese Medicine.

[0012] The fifth objective of this invention is to provide a computer-readable storage medium.

[0013] The sixth objective of this invention is to provide a computer program product.

[0014] To achieve the above objectives, the present invention is implemented through the following solution: This invention claims protection for the use of a group of biomarkers in the preparation of products for diagnosing and / or predicting dampness syndrome in Traditional Chinese Medicine, said biomarkers including body mass index, white blood cell count, neutrophil percentage, monocyte percentage, lactate dehydrogenase, high-density lipoprotein, apolipoprotein A, adiponectin, ε-lysine, pentosine, carboxyethyllysine, formyllysine, carboxymethyllysine, hydroxyimidazolinone, and / or acetylsine.

[0015] Preferably, the product is a detection reagent and / or a test kit.

[0016] The present invention also claims protection for a diagnostic reagent for diagnosing and / or predicting dampness syndrome in traditional Chinese medicine, the diagnostic reagent comprising products for detecting the aforementioned biomarkers.

[0017] The present invention also claims protection for a kit for diagnosing and / or predicting dampness syndrome in traditional Chinese medicine, the kit containing the above-described detection reagents.

[0018] This invention also claims protection for a product for diagnosing and / or predicting dampness syndrome in Traditional Chinese Medicine, containing a detection reagent and / or kit for detecting the aforementioned biomarkers.

[0019] This invention also claims protection for a machine learning model for predicting dampness syndrome in traditional Chinese medicine, including a data acquisition module, a data processing module, a prediction module, and a result output module; The data acquisition module is used to acquire the content of biomarkers in the sample to be tested; the biomarkers are those mentioned above. The data processing module takes the content of biomarkers in the sample to be tested obtained by the data acquisition module as input, and calculates the predicted score of the sample to be tested by combining it with Formula I. Formula I: Predicted score = β1 + β2 * Body mass index + β3 * White blood cell count + β4 * Neutrophil percentage + β5 * Monocyte percentage + β6 * Lactate dehydrogenase + β7 * High-density lipoprotein + β8 * Apolipoprotein A + β9 * Adiponectin + β 10 *Formyllysine + β 11 *Acetyllysine + β 12 *Carboxyethyl lysine + β 13 *ε-Lysine + β 14 *Carboxymethyl lysine + β 15 *Hydroxyimidazolinone + β 16 *Pentose; In Formula I, β1~β 16 Obtained through a regression model; The prediction module is based on the prediction score of the test sample obtained by the data processing module. The test sample with a prediction score ≥ 0.5 is judged as a TCM dampness syndrome sample, and the test sample with a prediction score < 0.5 is judged as a non-TCM dampness syndrome sample. The result output module is used to output the results obtained by the prediction module.

[0020] Preferably, in the data processing module, β1 to β in Formula I 16 It was obtained through the Lasso-logistics regression model.

[0021] More preferably, the regularization parameter λ of the Lasso-logistics regression model is 0.00564488.

[0022] More preferably, β1=-1.21, β2=0.179, β3=0.265, β4=0.034, β5=-0.047, β6=-0.021, β7=2.56, β8=-4.38, β9=-0.108, β in formula I 10 =0.001, β 11 =0.002, β 12 =0.032, β 13 =-0.477, β 14 =0.004, β 15 =0.012, β 16 =0.029.

[0023] Preferably, the TCM dampness syndrome samples in the prediction module are TCM dampness syndrome samples that meet the diagnostic criteria of the "Diagnostic Criteria for Dampness Syndrome" (T / CACM 1454-2023) and the "Classification and Determination of TCM Constitution".

[0024] This invention also claims protection for a prediction system for predicting dampness syndrome in Traditional Chinese Medicine, comprising an acquisition unit, a storage unit, and a processing unit, wherein the acquisition unit is used to acquire the content of biomarkers in a sample to be tested; the biomarkers are the aforementioned biomarkers; The storage unit stores program instructions that can be executed by the processing unit; The processing unit contains any of the machine learning models described above; When the program instructions are executed by the processing unit, the content of biomarkers in the sample to be tested obtained by the acquisition unit is input into the processing unit to obtain the TCM dampness syndrome prediction result.

[0025] The present invention also claims protection for a computer-readable storage medium having a computer program stored thereon that, when executed by a processor, implements any of the machine learning models described above.

[0026] Compared with the prior art, the present invention has the following beneficial effects: This invention provides a set of biomarkers for diagnosing and / or predicting dampness syndrome in Traditional Chinese Medicine (TCM). These biomarkers include body mass index (BMI), white blood cell count, neutrophil percentage, monocyte percentage, lactate dehydrogenase (LDH), high-density lipoprotein (HDL), apolipoprotein A (ALA), adiponectin, ε-lysine, pentosine, carboxyethyllysine, formyllysine, carboxymethyllysine, hydroxyimidazolinone, and / or acetyllysine. Based on these biomarkers, a machine learning model for predicting dampness syndrome in TCM is provided. By collecting the levels of these biomarkers in a sample and combining this data with Formula I, accurate diagnosis and / or prediction of dampness syndrome in TCM can be achieved.

[0027] When using the machine learning model to predict dampness syndrome in traditional Chinese medicine, the diagnosis and prediction are based on objectively detected biomarkers, which realizes the standardized judgment of the risk of dampness syndrome in traditional Chinese medicine, significantly improves the repeatability of diagnosis, and completely breaks through the subjective dependence barrier in the diagnosis of dampness syndrome in traditional Chinese medicine, avoiding the lack of diagnostic consistency caused by reliance on doctors' experience, limited by vague symptom descriptions and individual differences. Attached Figure Description

[0028] Figure 1 This is a graph showing the results of the Lasso-logistics regression analysis in Example 1; Figure 2 This is a graph showing the results of Lasso-logistics regression analysis using 10-fold cross-validation in Example 2; Figure 3 This is the ROC curve used in Example 2 to predict the dampness status of each patient in the training sample using the predictive scoring formula; Figure 4 This is the ROC curve diagram for predicting the dampness syndrome status of each patient in the test sample using Formula I in Example 2; Figure 5 This is the consistency analysis calibration chart from Example 2; Figure 6 This is the clinical decision curve from Example 2; Figure 7 ROC curves for predicting TCM dampness syndrome in patients in Comparative Example 1 by combining the BMI of each patient with the training sample. Figure 8 ROC curves for predicting the TCM dampness syndrome in each patient in Comparative Example 1 by combining the white blood cell count of each patient with the training sample. Figure 9 ROC curves for predicting the percentage of neutrophils in each patient in Comparative Example 1 when using training samples to diagnose dampness syndrome in traditional Chinese medicine. Figure 10 ROC curves for predicting the percentage of monocytes in each patient in Comparative Example 1 when combined with the training sample to indicate dampness syndrome in Traditional Chinese Medicine. Figure 11 ROC curves for predicting the TCM dampness syndrome in each patient in Comparative Example 1 based on the lactate dehydrogenase content of each patient in the training sample. Figure 12 ROC curves for predicting the TCM dampness syndrome in each patient in Comparative Example 1 by combining the high-density lipoprotein content of each patient with the training sample. Figure 13 ROC curves for predicting the TCM dampness syndrome in each patient in Comparative Example 1 by combining the apolipoprotein A content of each patient with the training sample. Figure 14 ROC curves for predicting the TCM dampness syndrome in each patient in Comparative Example 1 based on the adiponectin content of each patient in the training sample. Figure 15 ROC curves for predicting the TCM dampness syndrome in each patient in Comparative Example 1 based on the formyl lysine content of each patient in the training sample. Figure 16 ROC curves for predicting the TCM dampness syndrome in each patient in Comparative Example 1 based on the acetyllysine content of each patient in the training sample. Figure 17 ROC curves for predicting the TCM dampness syndrome in each patient in Comparative Example 1 based on the content of carboxyethyl lysine in each patient combined with the training samples. Figure 18ROC curves for predicting the TCM dampness syndrome in each patient in Comparative Example 1 by combining the ε-lysine content of each patient with the training sample. Figure 19 ROC curves for predicting the TCM dampness syndrome in each patient in Comparative Example 1 based on the carboxymethyl lysine content of each patient in the training sample. Figure 20 ROC curves for predicting the TCM dampness syndrome in each patient in Comparative Example 1 based on the hydroxyimidazolinone content of each patient in the training sample. Figure 21 The ROC curves for predicting the TCM dampness syndrome in each patient in Comparative Example 1 based on the pentasaccharide content of each patient in the training sample are shown. Detailed Implementation

[0029] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. These embodiments are for illustrative purposes only and are not intended to limit the scope of the invention. Unless otherwise specified, the experimental methods used in the following embodiments are conventional methods; the materials and reagents used, unless otherwise specified, are commercially available.

[0030] All clinical patient sample data used in the embodiments of this invention have been accompanied by informed consent and authorization from the patients.

[0031] Example 1: Determination of biomarkers related to dampness syndrome in Traditional Chinese Medicine I. Determination of Differential Samples 1. Experimental Methods A total of 634 patients who underwent physical examinations at the Guangzhou Municipal Cadre and Talent Health Management Center from January 2023 to May 2024 were selected. Referring to the "Diagnostic Criteria for Dampness Syndrome" (T / CACM 1454-2023) and the "Classification and Determination of Traditional Chinese Medicine Constitution," the 634 patients were divided into a balanced constitution group (containing 308 balanced constitution patients) and a dampness syndrome group (containing 326 TCM dampness syndrome patients). The selection criteria for the 634 patients were: exclusion of patients with severe heart, brain, or kidney dysfunction, impaired consciousness, mental abnormalities, inability to complete the questionnaire, severe systemic infections, tumor patients, and pregnant or expectant women.

[0032] The 634 patients were randomly divided into training and testing samples at a 1:1 ratio. The training and testing samples included 154 patients with dampness syndrome according to traditional Chinese medicine and 163 patients with balanced constitution.

[0033] For patients in the balanced constitution group and the dampness syndrome group, a fully automated hematology analyzer (BC-6800Plus, Shenzhen Mindray Bio-Medical Electronics Co., Ltd.) was used to test the complete blood count (including red blood cell count, white blood cell count, platelet count, hemoglobin concentration, mean corpuscular volume, mean platelet volume, neutrophil percentage, lymphocyte percentage, monocyte percentage, eosinophil percentage, and basophil percentage); a fully automated biochemical and immunoassay analyzer (Cobas) was used. Serum biochemical parameters (including alanine aminotransferase, aspartate aminotransferase, total bilirubin, indirect bilirubin, total protein, albumin, globulin, albumin-globulin ratio, alkaline phosphatase, gamma-glutamyl transferase, lactate dehydrogenase, bile acids, creatinine, fasting blood glucose, total cholesterol, triglycerides, low-density lipoprotein, high-density lipoprotein, apolipoprotein A, apolipoprotein B, uric acid, and adiponectin) were measured in each patient using a Waters Xevo TQ-S triple quadrupole mass spectrometer coupled with an ACQUITY UPLC™ system, according to existing technology ("Development of simultaneous quantitation method for 20 free advanced glycation end products using UPLC-MS / MS and clinical application in kidney injury" (DOI: 800-c702, Beijing Wantai Derui Diagnostic Technology Co., Ltd.). Mass spectrometry analysis was performed to detect the content of advanced glycosylation end products (including formyl lysine, carboxymethyl cysteine, acetyl lysine, carboxyethyl lysine, ε-(2-formyl-5-hydroxymethylpyrrole-1-yl)-lysine, carboxymethyl lysine, hydroxyimidazolinone, and pentose glycosides) as shown in 10.1016 / j.jpba.2024.116035; PMID: 38367518).

[0034] Next, for patients in the balanced constitution group and the dampness syndrome group, the T-test was used to analyze the differences in blood routine indicators, serum biochemical indicators and advanced glycation end products content between the two groups (P < 0.05 was considered significant).

[0035] 2. Experimental Results Table 1 shows the biomarkers with significant differences in the T-test results for patients in the balanced constitution group and the dampness syndrome group.

[0036] Table 1. Biomarkers showing significant differences in the T-test results.

[0037] The results showed that when t-tests were performed on the levels of various biomarkers in the balanced constitution group and the TCM dampness syndrome group, there were differences in the biomarkers between the two groups, indicating that the division between the balanced constitution group and the TCM dampness syndrome group was meaningful.

[0038] II. Determination of biomarkers related to dampness syndrome in Traditional Chinese Medicine 1. Experimental Methods Using the training samples from step one as the object, the biomarkers (including blood routine indicators, serum biochemical indicators and the content of advanced glycosylation end products) in the training samples were obtained as shown in step one. The glmnet package was used to perform a one-way competitive risk analysis on each biomarker in the training samples in combination with the actual wet syndrome status (wet syndrome or balanced constitution) of the training samples, and biomarkers with statistical differences (P < 0.1) were screened out.

[0039] Next, for the biomarkers with statistical differences obtained from the univariate regression analysis, Lasso-logistics regression analysis was performed on the training samples in combination with the actual wet syndrome status (wet syndrome or balanced constitution).

[0040] 2. Experimental Results The results of the univariate competitive risk analysis and Lasso-logistics regression analysis are shown in Table 2; the results of the Lasso-logistics regression analysis are shown in the graph. Figure 1 As shown.

[0041] Table 2 Results of univariate competition risk analysis and Lasso-logistics regression analysis

[0042] The results showed that, based on univariate regression analysis, 15 biomarkers were significantly different between patients with dampness syndrome and those with peaceful constitution in the training sample. These biomarkers were: body mass index (BMI), white blood cell count (WBC), neutrophil percentage (NEUT%), monocyte percentage (MONO%), lactate dehydrogenase (LDH), high-density lipoprotein (HDL), apolipoprotein A (APOA1), adiponectin (ADPN), ε-lysine, pentosine, carboxyethyllysine, formyllysine (N6-FL), carboxymethyllysine, hydroxyimidazolinone, and acetyllysine.

[0043] Example 2: A predictive scoring formula for diagnosing dampness syndrome in Traditional Chinese Medicine I. Training of Predictive Scoring Formula 1. Experimental Methods The levels of 15 TCM dampness-related biomarkers in each patient in the training sample of Example 1 (317 cases, including 154 TCM dampness syndrome patients and 163 balanced constitution patients) were determined according to the method shown in Example 1; wherein the 15 TCM dampness-related biomarkers are as shown in Example 1.

[0044] Next, the content of TCM dampness-related biomarkers in each patient in the training sample was used as a feature, and the dampness state (dampness syndrome or balanced constitution) of each patient in the training sample was used as the prediction target. Lasso-logistics regression analysis was performed using the training sample through 10-fold cross-validation to obtain the optimal penalty coefficient λ of Lasso-logistics regression analysis, and the prediction score formula was obtained.

[0045] Samples with a predicted score ≥ 0.5 (judgment threshold) calculated based on the prediction scoring formula are denoted as TCM dampness syndrome samples, and samples with a predicted score < 0.5 calculated based on the prediction scoring formula are denoted as non-TCM dampness syndrome samples (balanced constitution samples).

[0046] Next, the expression levels of TCM dampness-related biomarkers in each patient in the training sample were combined with the prediction scoring formula to determine the dampness status of each patient in the training sample. Then, based on the actual dampness status of each patient in the training sample, the AUC, sensitivity, and specificity of predicting TCM dampness status of each patient in the training sample using the prediction scoring formula were calculated.

[0047] 2. Experimental Results The results of Lasso-logistic regression analysis using 10-fold cross-validation based on the expression levels of TCM-related biomarkers for dampness syndrome in each patient in the training sample and the dampness syndrome status are shown in the figure below. Figure 2 As shown in the figure, the optimal penalty coefficient λ for Lasso-logistics regression analysis is 0.00564488.

[0048] The predicted score formula obtained from training is shown in Formula I.

[0049] Formula I: Predicted score = -1.21 + 0.179 × Body mass index + 0.265 × White blood cell count + 0.034 × Neutrophil percentage - 0.047 × Monocyte percentage - 0.021 × Lactate dehydrogenase + 2.56 × High-density lipoprotein - 4.38 × Apolipoprotein A - 0.108 × Adiponectin + 0.001 × Formyl lysine + 0.002 × Acetyl lysine + 0.032 × Carboxyethyl lysine - 0.477 × ε-lysine + 0.004 × Carboxymethyl lysine + 0.012 × Hydroxyimidazolinone + 0.029 × Pentosine.

[0050] The ROC curve for predicting the dampness syndrome status of each patient in the training sample using the predictive scoring formula (Formula I) is as follows: Figure 3 As shown, the area under the ROC curve (AUC) is 0.803, the sensitivity is 0.740, and the specificity is 0.730.

[0051] This demonstrates that the predictive scoring formula (Formula I) can effectively distinguish between patients with dampness syndrome in Traditional Chinese Medicine (TCM) and those without dampness syndrome (patients with a balanced constitution), and can effectively predict the dampness syndrome status of the sample.

[0052] II. Testing the Predictive Scoring Formula 1. Experimental Methods The levels of 15 TCM dampness-related biomarkers in each patient in the test sample of Example 1 (317 cases, including 154 TCM dampness syndrome patients and 163 balanced constitution patients) were determined according to the method shown in Example 1; wherein the 15 TCM dampness-related biomarkers are as shown in Example 1.

[0053] Next, the levels of TCM dampness-related biomarkers in each patient in the test sample are combined with the prediction scoring formula shown in Formula I to determine the dampness status of each patient in the test sample. Based on the actual dampness status of each patient in the test sample, the AUC, sensitivity, and specificity of predicting TCM dampness in each patient in the test sample using the prediction scoring formula shown in Formula I are calculated.

[0054] 2. Experimental Methods The ROC curve for predicting the dampness syndrome status of each patient in the test sample using the prediction scoring formula shown in Formula I is as follows: Figure 4 As shown, the area under the ROC curve (AUC) is 0.801, the sensitivity is 712, and the specificity is 0.728.

[0055] The prediction scoring formula shown in Formula I demonstrates excellent discriminative ability in predicting the dampness syndrome status of patients in independent test samples. The formula exhibits excellent generalization ability and can accurately determine the dampness syndrome status of samples in different sample data.

[0056] III. Efficacy Evaluation and Clinical Decision Analysis of Predictive Scoring Formulas 1. Experimental Methods The levels of 15 TCM dampness-related biomarkers in each patient in the training sample of Example 1 (317 cases, including 154 TCM dampness syndrome patients and 163 balanced constitution patients) were determined according to the method shown in Example 1; wherein the 15 TCM dampness-related biomarkers are as shown in Example 1.

[0057] Next, the levels of 15 TCM dampness syndrome-related biomarkers in each patient in the training sample were combined with the prediction scoring formula shown in Formula I to calculate the predicted score for each patient in the training sample. The patients were then sorted from low to high according to the predicted scores. The training sample was then divided into 10 groups using the decimal grouping method. The average predicted score and the actual proportion of TCM dampness syndrome patients in each group were calculated. Consistency analysis was performed based on the average predicted score and the actual proportion of TCM dampness syndrome patients, and a calibration curve was plotted. At the same time, the Brier score (mean squared error) between the average predicted score and the actual proportion of TCM dampness syndrome patients was calculated.

[0058] Next, based on the predicted scores of each patient in the training samples obtained by calculation, the judgment thresholds were set to 0.01 to 0.99 respectively. With "all patients treated" (all samples are regarded as patients with dampness syndrome in TCM, true positives are actual patients with dampness syndrome in TCM, and false positives are actual patients with balanced constitution) and "no treatment" (all samples are regarded as patients without dampness syndrome in TCM, true positives = 0, false positives = 0) as references, clinical decision analysis was carried out and clinical decision analysis curves were plotted.

[0059] 2. Experimental Results Consistency analysis calibration curves are shown below Figure 5 As shown in the clinical decision curve diagram... Figure 6 As shown.

[0060] In the consistency analysis calibration curve, the actual calibration curve is close to the ideal calibration curve. The Brier score is 0.1813, and the slope of the actual calibration curve is 0.99. Furthermore, the clinical decision curve shows that when using the prediction scoring formula shown in Formula I to predict the wet syndrome status of the sample, the net benefit of clinical decision is significantly higher when the judgment threshold is between 0.3 and 0.8, and 0.5 is the optimal judgment threshold.

[0061] The prediction scoring formula shown in Formula I, when combined with the judgment threshold to judge the TCM dampness syndrome status of the samples, has high discriminative power, consistency and clinical utility.

[0062] Example 3: A machine learning model for predicting dampness syndrome in Traditional Chinese Medicine A machine learning model for predicting dampness syndrome in traditional Chinese medicine includes a data acquisition module, a data processing module, a prediction module, and a result output module. The data acquisition module is used to obtain the content of 15 TCM-related biomarkers for dampness syndrome in the sample to be tested; The 15 biomarkers related to dampness syndrome in Traditional Chinese Medicine are: body mass index (BMI), white blood cell count (WBC), neutrophil percentage (NEUT%), monocyte percentage (MONO%), lactate dehydrogenase (LDH), high-density lipoprotein (HDL), apolipoprotein A (APOA1), adiponectin (ADPN), ε-lysine, pentosine, carboxyethyllysine, formyllysine (N6-FL), carboxymethyllysine, hydroxyimidazolinone, and acetyllysine; The data processing module takes the content of 15 TCM dampness-related biomarkers in the sample to be tested obtained by the data acquisition module as input, and calculates the predicted score of the sample to be tested by combining the predicted score formula shown in Formula I obtained in Example 2. The prediction module is based on the prediction score of the test sample obtained by the data processing module. The test sample with a prediction score ≥ 0.5 is judged as a TCM dampness syndrome sample, and the test sample with a prediction score < 0.5 is judged as a non-TCM dampness syndrome sample. The result output module is used to output the results obtained by the prediction module.

[0063] Comparative Example 1: A predictive scoring formula for diagnosing dampness syndrome in Traditional Chinese Medicine I. Experimental Methods The expression levels of 15 different indicators were measured in each patient in the training sample of Example 2, including body mass index (BMI), white blood cell count (WBC), neutrophil percentage (NEUT%), monocyte percentage (MONO%), lactate dehydrogenase (LDH), high-density lipoprotein (HDL), apolipoprotein A (APOA1), adiponectin (ADPN), formyl lysine (N6-FL), acetyl lysine, carboxyethyl lysine, ε-lysine, carboxymethyl lysine, hydroxyimidazolinone, and pentosine.

[0064] Next, following the method shown in step one of Example 2, the expression level of each indicator in each patient in the training sample and the dampness syndrome status (dampness syndrome or balanced constitution) of each patient in the training sample are combined as prediction targets to construct a prediction scoring formula and draw the ROC curve when predicting the TCM dampness syndrome of each patient in the training sample using the prediction scoring formula, and calculate the corresponding AUC value.

[0065] II. Experimental Results The ROC curve for predicting the TCM dampness syndrome of each patient based on their BMI in the training sample is shown below. Figure 7 As shown; the ROC curve for predicting the TCM dampness syndrome of each patient by combining the white blood cell count of each patient in the training sample is shown in the figure. Figure 8 As shown; the ROC curve for predicting the TCM dampness syndrome of each patient by combining the neutrophil percentage (NEUT%) of each patient in the training sample is shown in the figure. Figure 9As shown; the ROC curve for predicting the TCM dampness syndrome of each patient by combining the percentage of monocytes (MONO%) in each patient's training sample is shown in the figure. Figure 10 As shown; the ROC curve for predicting the TCM dampness syndrome of each patient by combining the lactate dehydrogenase (LDH) content of each patient in the training sample is shown in the figure. Figure 11 As shown; the ROC curve for predicting the TCM dampness syndrome of each patient by combining the high-density lipoprotein (HDL) content of each patient in the training sample is shown in the figure. Figure 12 As shown; the ROC curve for predicting the TCM dampness syndrome of each patient by combining the apolipoprotein A (APOA1) content of each patient in the training sample is shown in the figure. Figure 13 As shown; the ROC curve for predicting the TCM dampness syndrome of each patient by combining the adiponectin (ADPN) content of each patient in the training sample is shown in the figure. Figure 14 As shown; the ROC curve for predicting the TCM dampness syndrome of each patient by combining the formyl lysine (N6-FL) content of each patient in the training sample is shown in the figure. Figure 15 As shown; the ROC curve for predicting the TCM dampness syndrome of each patient by combining the acetyl-lysine content of each patient in the training sample is shown in the figure. Figure 16 As shown; the ROC curve for predicting the TCM dampness syndrome of each patient based on the carboxyethyl lysine content in the training sample is shown below. Figure 17 As shown; the ROC curve for predicting the TCM dampness syndrome of each patient by combining the ε-lysine content of each patient in the training sample is shown in the figure. Figure 18 As shown; the ROC curve for predicting the TCM dampness syndrome of each patient based on the carboxymethyl lysine content of each patient in the training sample is shown in the figure. Figure 19 As shown; the ROC curve for predicting the TCM dampness syndrome of each patient based on the hydroxyimidazolinone content in the training sample is shown in the figure. Figure 20 As shown; the ROC curve for predicting the TCM dampness syndrome of each patient by combining the pentasaccharide content of each patient in the training sample is shown in the figure. Figure 21 As shown.

[0066] The results showed that when the prediction scoring formula combining the expression level of a single indicator in the training samples was used to predict the TCM dampness syndrome of patients, the area under the ROC curve was <0.8 (basically between 0.55 and 0.60), indicating that combining the expression level of a single indicator could not achieve accurate prediction of TCM dampness syndrome of patients.

[0067] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the scope of protection of the present invention. For those skilled in the art, other variations or modifications can be made based on the above description and ideas, and it is neither necessary nor possible to exhaustively describe all implementation methods here. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the claims of the present invention.

Claims

1. The application of a group of biomarkers in the preparation of products for diagnosing and / or predicting dampness syndrome in Traditional Chinese Medicine, characterized in that, The biomarkers include body mass index, white blood cell count, neutrophil percentage, monocyte percentage, lactate dehydrogenase, high-density lipoprotein, apolipoprotein A, adiponectin, ε-lysine, pentosine, carboxyethyllysine, formyllysine, carboxymethyllysine, hydroxyimidazolinone, and acetyllysine.

2. The application according to claim 1, characterized in that, The products mentioned are testing reagents and / or kits.

3. A product for diagnosing and / or predicting dampness syndrome in Traditional Chinese Medicine, characterized in that, Detection reagents and / or kits containing the biomarkers described in claim 1.

4. A machine learning model for predicting dampness syndrome in Traditional Chinese Medicine, characterized in that, It includes a data acquisition module, a data processing module, a prediction module, and a result output module; The data acquisition module is used to acquire the content of biomarkers in the sample to be tested; the biomarkers are those described in claim 1. The data processing module takes the content of biomarkers in the sample to be tested obtained by the data acquisition module as input, and calculates the predicted score of the sample to be tested by combining it with Formula I. Formula I: Predicted score = β1 + β2 * Body mass index + β3 * White blood cell count + β4 * Neutrophil percentage + β5 * Monocyte percentage + β6 * Lactate dehydrogenase + β7 * High-density lipoprotein + β8 * Apolipoprotein A + β9 * Adiponectin + β 10 *Formyllysine + β 11 *Acetyllysine + β 12 *Carboxyethyl lysine + β 13 *ε-Lysine + β 14 *Carboxymethyl lysine + β 15 *Hydroxyimidazolinone + β 16 *Pentose; In Formula I, β1~β 16 Obtained through a regression model; The prediction module is based on the prediction score of the test sample obtained by the data processing module. The test sample with a prediction score ≥ 0.5 is judged as a TCM dampness syndrome sample, and the test sample with a prediction score < 0.5 is judged as a non-TCM dampness syndrome sample. The result output module is used to output the results obtained by the prediction module.

5. The machine learning model according to claim 4, characterized in that, In the data processing module, β1~β in Formula I 16 It was obtained through the Lasso-logistics regression model.

6. The machine learning model according to claim 5, characterized in that, The regularization parameter λ of the Lasso-logistics regression model is 0.00564488.

7. The machine learning model according to claim 4, characterized in that, In Formula I, β1=-1.21, β2=0.179, β3=0.265, β4=0.034, β5=-0.047, β6=-0.021, β7=2.56, β8=-4.38, β9=-0.108, β 10 =0.001,β 11 =0.002,β 12 =0.032,β 13 =-0.477,β 14 =0.004,β 15 =0.012,β 16 =0.

029.

8. A prediction system for predicting dampness syndrome in Traditional Chinese Medicine, comprising an acquisition unit, a storage unit, and a processing unit, characterized in that, The acquisition unit is used to acquire the content of biomarkers in the sample to be tested; the biomarker is the biomarker described in claim 1; The storage unit stores program instructions that can be executed by the processing unit; The processing unit contains the machine learning model according to any one of claims 4 to 7; When the program instructions are executed by the processing unit, the content of biomarkers in the sample to be tested obtained by the acquisition unit is input into the processing unit to obtain the TCM dampness syndrome prediction result.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements any one of the machine learning models described in claims 4 to 7.

10. A computer program product, characterized in that, The computer program product includes the computer-readable storage medium of claim 9.

Images