Method for constructing a model for predicting the risk of metabolic syndrome due to antipsychotic drugs and system thereof

Abstract

The application relates to the technical field of medical data processing, and discloses a construction method and system of a model for predicting the risk of metabolic syndrome caused by antipsychotic drugs, which comprises the following steps: acquiring clinical information and metabolic index data of a discovery cohort and a verification cohort respectively; acquiring genome data of the discovery cohort, screening significant SNP sites through whole genome association analysis, and calculating a polygenic risk score; acquiring proteome data of the verification cohort, screening candidate proteins through metabolic outcome association analysis; taking PRS data and / or candidate protein level data as independent variables, and taking the percentage change of metabolic indicators as a risk label to train a prediction model. The application integrates multi-omics information of genomes and proteomes to construct a prediction model capable of quantitatively evaluating the risk of metabolic syndrome caused by antipsychotic drugs for individuals, which can assist doctors in accurately identifying high-risk patients before or in the early stage, and provides an important technical tool for formulating individualized treatment plans and reducing the risk of MetS in clinical practice.

Description

Technical Field

[0001] This invention relates to the field of medical data processing technology, and in particular to a method and system for constructing a model for predicting the risk of metabolic syndrome caused by antipsychotic drugs. Background Technology

[0002] Schizophrenia (SCZ) is a severe mental disorder with a high disease burden and disability rate. Antipsychotics (APDs) are the first-line treatment, but they significantly increase the risk of metabolic syndrome (MetS), with an estimated prevalence of 22.4%–63%, two to three times higher than in the general population. Metabolic syndrome typically includes abdominal obesity, insulin resistance, hypertension, elevated triglycerides, and decreased high-density lipoprotein cholesterol. Patients with APS-induced MetS have a significantly increased risk of developing type 2 diabetes, cardiovascular disease, and other metabolic-related diseases, which are leading causes of death in schizophrenia patients and severely impact their long-term prognosis.

[0003] Currently, the prediction of the risk of metabolic syndrome caused by antipsychotic drugs in clinical practice mainly relies on an experience-based, manual model. This model involves physicians subjectively integrating and weighing the known metabolic side effects of the drug with the patient's baseline characteristics based on their professional knowledge, resulting in a vague risk assessment. This method has significant drawbacks: First, it is highly subjective, heavily reliant on the physician's personal experience, leading to poor consistency among different assessors; second, it is inefficient, struggling to systematically and rapidly process a large number of multi-dimensional risk factors; and finally, this crude assessment has limited accuracy, failing to perform individualized quantitative risk stratification, often resulting in delayed warnings and missed opportunities for early prevention.

[0004] Therefore, there is an urgent need to develop a model for predicting the risk of metabolic syndrome caused by antipsychotic drugs. By quantitatively assessing the risk of patients developing metabolic syndrome before prescription, this model can assist doctors in developing reasonable individualized treatment plans for patients, thereby improving the safety of treatment and the level of long-term health management for patients. Summary of the Invention

[0005] This invention provides a method and system for constructing a model for predicting the risk of metabolic syndrome caused by antipsychotic drugs, thereby overcoming the deficiencies of the prior art.

[0006] This invention provides a method for constructing a model for predicting the risk of metabolic syndrome induced by antipsychotic drugs, comprising: Clinical information and metabolic data were obtained from the discovery cohort and the validation cohort, respectively. The discovery cohort consisted of schizophrenia patients receiving different antipsychotic drug treatments, while the validation cohort consisted of schizophrenia patients receiving uniform antipsychotic drug treatments. Both the discovery cohort and the validation cohort included patients who developed new-onset antipsychotic drug-induced metabolic syndrome within a set treatment period. Genomic data from the discovery and validation cohorts were obtained, and combined with their clinical information and metabolic marker data, genome-wide association analysis was performed on the discovery cohort to screen for significant SNP sites that are significantly associated with antipsychotic drug-induced metabolic syndrome. Based on this, polygenic risk score data for the validation cohort were obtained. Proteomic data of the validation cohort were obtained, significant SNP sites were located to genes to obtain candidate proteins, and metabolic outcome association analysis was performed by combining the protein level data of the candidate proteins with their metabolic index data to obtain the protein level data of the target proteins associated with antipsychotic drug-induced metabolic syndrome in the validation cohort. Using multigene risk score data from the validation cohort and / or protein level data of the target protein from the validation cohort as independent variables, and the percentage change of metabolic indicators from the baseline within a set treatment period from the validation cohort as risk labels, a model for predicting the risk of metabolic syndrome induced by antipsychotic drugs was trained.

[0007] According to the present invention, a method for constructing a model for predicting the risk of metabolic syndrome caused by antipsychotic drugs is provided. The clinical information includes any one or any combination of the following: demographic information (age, sex, disease duration, etc.), height and weight, waist circumference, type of antipsychotic drug, and dosage of antipsychotic drug.

[0008] According to the present invention, a method for constructing a model for predicting the risk of metabolic syndrome caused by antipsychotic drugs is provided, wherein the metabolic indicators include fasting blood glucose and / or blood lipids (total cholesterol, triglycerides, high-density lipoprotein, low-density lipoprotein).

[0009] According to the present invention, a method for constructing a model for predicting the risk of metabolic syndrome induced by antipsychotic drugs is provided. The process involves acquiring genomic data from the discovery and validation cohorts, combining this data with clinical information and metabolic marker data, performing genome-wide association analysis on the discovery cohort, screening for significant SNPs associated with antipsychotic drug-induced metabolic syndrome, and obtaining polygenic risk score data for the validation cohort based on this. Preprocess the genomic data; Based on the genomic data of the discovery cohort, the genetic principal component data of the discovery cohort were obtained; Using the discovery of new-onset antipsychotic drug-induced metabolic syndrome within a set treatment period as the endpoint, clinical data and principal component data of the discovery cohort were used as covariates to construct a logistic regression model, and genome-wide association analysis was performed to screen out significant SNP sites that are significantly associated with antipsychotic drug-induced metabolic syndrome. Based on the results of genome-wide association analysis, significant SNP sites were located to genes, and a multigene risk score was constructed in the validation cohort. Sets of significant SNP sites under different P-value thresholds were screened, and the set of significant SNP sites under the optimal P-value threshold was obtained through fitting. The multigene risk score data of the validation cohort under the set of significant SNP sites corresponding to the optimal P-value threshold were obtained.

[0010] According to the present invention, a method for constructing a model for predicting the risk of metabolic syndrome caused by antipsychotic drugs is provided. The set of significant SNP sites includes the intergenic SNP rs73762168 located at 6q21, or genes (e.g., NR2E1, SNX3, and AFG1L / LACE1) located near rs73762168.

[0011] According to the present invention, a method for constructing a model for predicting the risk of metabolic syndrome induced by antipsychotic drugs is provided. The process involves acquiring proteomic data from the validation cohort, locating significant SNP sites to genes to obtain candidate proteins, and combining the protein level data of the candidate proteins with their metabolic index data to perform metabolic outcome association analysis. This yields protein level data of target proteins in the validation cohort associated with antipsychotic drug-induced metabolic syndrome, including: Preprocess the proteomic data; Significant SNP sites are located to genes to obtain candidate proteins; Using the percentage change of metabolic parameters relative to baseline in the validation cohort over a set treatment period as the outcome, and clinical information from the validation cohort as covariates, a multiple linear regression model was constructed to conduct an association analysis of metabolic outcomes. This identified target proteins associated with the risk of antipsychotic drug-induced metabolic syndrome and obtained protein level data of target proteins associated with antipsychotic drug-induced metabolic syndrome in the validation cohort.

[0012] According to the present invention, a method for constructing a model for predicting the risk of metabolic syndrome caused by antipsychotic drugs is provided, wherein the target protein is SNX3 protein.

[0013] According to the present invention, a method for constructing a model for predicting the risk of metabolic syndrome induced by antipsychotic drugs is provided. The model, trained using multigene risk score data from the validation cohort and / or protein level data of the target protein from the validation cohort as independent variables, and the percentage change of metabolic indicators from baseline within a set treatment period from the validation cohort as a risk label, for predicting the risk of metabolic syndrome induced by antipsychotic drugs, includes: The first prediction model was trained using the multigene risk score data of the validation cohort as the independent variable and the percentage change of the metabolic index data of the validation cohort relative to the baseline as the dependent variable. The second prediction model was trained by using the protein level data of the target protein in the validation cohort as the independent variable and the percentage change of the metabolic index data of the validation cohort relative to the baseline as the dependent variable. By combining multi-gene risk score data from the validation cohort and protein level data of the target protein as independent variables, and using the percentage change of metabolic index data from the validation cohort relative to the baseline as the risk label, a third prediction model was trained. The predictive performance of the first, second, and third prediction models was compared to obtain the preferred prediction model, which was then used as the final model for predicting the risk of metabolic syndrome caused by antipsychotic drugs.

[0014] According to the present invention, a method for constructing a model for predicting the risk of metabolic syndrome caused by antipsychotic drugs is provided, and the final model for predicting the risk of metabolic syndrome caused by antipsychotic drugs is a third prediction model.

[0015] This invention also provides a risk prediction system for metabolic syndrome induced by antipsychotic drugs, comprising: The test data acquisition module is used to receive clinical data, metabolic index data, genomic data and proteomic data of the test subject from at least one terminal, wherein the test subject is a schizophrenic patient who is preparing to receive a certain antipsychotic drug treatment plan; The data processing module is used to: according to the method for constructing a model for predicting the risk of metabolic syndrome caused by antipsychotic drugs as described in any of the above, obtain the multigene risk score data of the subject under the set of significant SNP sites corresponding to the optimal P-value threshold based on the subject's genomic data; The prediction module is used to: based on the multigene risk score data of the subject under the set of significant SNP sites corresponding to the optimal P-value threshold, as well as clinical data, metabolic index data, and protein data, and through a model used to predict the risk of metabolic syndrome caused by antipsychotic drugs, obtain the prediction result of the subject's risk of metabolic syndrome caused by a certain antipsychotic drug treatment regimen. The prediction result output module is used to output the prediction results of the risk of metabolic syndrome caused by antipsychotic drugs under a certain antipsychotic drug treatment regimen to at least one terminal.

[0016] It should be noted that a terminal refers to an input / output device connected to a computer system. Depending on the function, terminals can be divided into various types: smart terminals or intelligent terminals, dumb terminals, interactive terminals or online terminals. Specifically, a terminal can be various mobile communication devices, such as mobile phones and tablets. This article aims to provide users with the function of inputting data and outputting data.

[0017] The present invention also provides an electronic device, including a processor and a memory storing a computer program, wherein the processor executes the computer program to implement the method for constructing a model for predicting the risk of metabolic syndrome induced by antipsychotic drugs as described above.

[0018] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method for constructing a model for predicting the risk of metabolic syndrome induced by antipsychotic drugs as described above.

[0019] The present invention also provides a computer program product comprising a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer is able to execute any of the above-described methods for constructing a model for predicting the risk of metabolic syndrome induced by antipsychotic drugs.

[0020] The present invention provides a method and system for constructing a model for predicting the risk of metabolic syndrome induced by antipsychotic drugs, which can bring at least the following beneficial effects: Based on a large sample of schizophrenia cases in China, this invention proposes and validates a multi-omics risk prediction model construction method that simultaneously integrates genomic and proteomic information. This method, through quantitative analysis of a polygenic risk score (PRS) reflecting an individual's genetic susceptibility and baseline plasma protein levels reflecting an individual's real-time functional status, can objectively and efficiently quantitatively assess the risk of developing metabolic syndrome under specific antipsychotic drug treatment. This changes the current clinical practice model, which mainly relies on physicians' subjective experience for vague risk assessments. It solves the problems of existing methods being highly subjective, inconsistent, inefficient, and having limited accuracy, and failing to achieve individualized quantitative risk stratification, significantly improving the objectivity, consistency, and accuracy of predictions.

[0021] This invention constructs a predictive model (the third predictive model) that integrates multi-omics information by jointly modeling genomic information (such as based on a set of significant SNP sites, particularly the intergenic SNP rs73762168 located in the 6q21 region and its related genes) and proteomic information (such as the target protein SNX3). This model fully utilizes complementary information from genetic background and physiological functional status, significantly improving the predictive accuracy of key metabolic phenotypes such as antipsychotic drug-related weight gain and waist circumference increase. This allows for high-precision identification of individuals at high risk of metabolic syndrome before or early in the course of medication, thus significantly advancing the early warning point and gaining valuable time for early preventive intervention.

[0022] Based on the model and system developed in this invention, clinicians can conveniently obtain patients' genotyping and baseline serum protein test data before prescribing antipsychotic drugs for schizophrenia patients, and quickly calculate individualized multi-omics risk scores. This score can assist physicians in: (a) selecting drug types or starting doses with lower metabolic risk for specific patients; (b) developing more conservative and individualized dosage adjustment strategies; and (c) for identified high-risk patients, developing and strengthening intensive monitoring plans and early intervention measures (such as lifestyle interventions and preventative medications) targeting metabolic indicators such as weight, waist circumference, blood glucose, and blood lipids. This can minimize the risk of metabolic syndrome while ensuring the core efficacy of antipsychotic treatment.

[0023] Through the aforementioned precise prediction and individualized management, this invention can reduce or delay the occurrence of metabolic syndrome and its subsequent complications (such as type 2 diabetes and cardiovascular disease) caused by antipsychotic drugs from the source. This not only directly improves patient safety during treatment and enhances patient tolerance and long-term adherence to drug therapy, but also helps improve patients' overall health and long-term prognosis, reduces all-cause mortality, and thus improves quality of life.

[0024] This invention can be easily integrated with hospital electronic medical record (EMR) systems and laboratory information management systems (LIS) to form semi-automated or even automated risk prediction and clinical decision support tools. This will further simplify clinical operation procedures, improve assessment efficiency, and provide solid technical support for achieving large-scale precision medication and individualized follow-up management in the field of psychiatry, demonstrating promising prospects for clinical translation and application. Attached Figure Description

[0025] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0026] Figure 1 This is a flowchart illustrating a method for constructing a model to predict the risk of metabolic syndrome caused by antipsychotic drugs, as provided by the present invention.

[0027] Figure 2 This is a schematic diagram of the structure of a system for constructing a model for predicting the risk of metabolic syndrome caused by antipsychotic drugs, as provided by the present invention.

[0028] Figure 3 This is a schematic diagram of the structure of the electronic device provided by the present invention. Detailed Implementation

[0029] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, embodiments of this invention, and should not be construed as limiting the invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention. In the description of this invention, it should be understood that the terminology used is for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0030] Terminology Explanation: 1. SCZ: Schizophrenia 2. APDs: Antipsychotics 3. MetS: Metabolic syndrome 4. SNP: Single Nucleotide Polymorphism 5. PRS: Polygenic Risk Score Figure 1This is a flowchart illustrating a method for constructing a model to predict the risk of metabolic syndrome caused by antipsychotic drugs, as provided by the present invention. The execution entity of this method can be any applicable terminal-side device or network-side device, such as a device for constructing a model to predict the risk of metabolic syndrome caused by antipsychotic drugs.

[0031] See Figure 1 The present invention provides a method for constructing a model for predicting the risk of metabolic syndrome induced by antipsychotic drugs, which may include: S110. Obtain clinical information and metabolic index data for the discovery cohort and the validation cohort, respectively. The discovery cohort consists of schizophrenia patients receiving different antipsychotic drug treatments, while the validation cohort consists of schizophrenia patients receiving uniform antipsychotic drug treatments. Both the discovery and validation cohorts include patients who develop new-onset antipsychotic drug-induced metabolic syndrome within a set treatment period. Clinical information includes any one or any combination of the following: demographic information (age, sex, disease duration, etc.), height and weight, waist circumference, type of antipsychotic drug, and dosage of antipsychotic drug. Metabolic indexes include fasting blood glucose and / or blood lipids (total cholesterol, triglycerides, high-density lipoprotein, and low-density lipoprotein).

[0032] In this embodiment, the cohort was derived from the Chinese Consortium for Pharmacogenomics of Antipsychotic Drugs (CAPOC). The SINO study, conducted in multiple psychiatric hospitals across China, included 3067 patients with schizophrenia. All participants were aged 18–45 years, diagnosed with schizophrenia via a DSM-IV structured clinical interview, hospitalized for acute exacerbations, with a PANSS total score ≥60 and at least 3 positive items ≥4 at enrollment, and generally normal physical examination and laboratory indicators, suitable for oral medication treatment. Exclusion criteria included: diagnosis of other mental disorders, presence of severe unstable physical illness, recent electroconvulsive therapy, need for long-acting injections for maintenance, significant liver or kidney dysfunction, pregnancy / breastfeeding, or contraindications to the investigational drug.

[0033] Subjects were randomly assigned to seven antipsychotic drug treatment groups (olanzapine, risperidone, quetiapine, aripiprazole, ziprasidone, perphenazine, and haloperidol). For the first two weeks, drug dosages were individually adjusted within the specified range, after which the dosages remained relatively stable. Throughout the 6-week follow-up period, demographic information (age, sex, disease duration, etc.), height and weight, waist circumference, type and dosage of antipsychotic drugs were systematically recorded. Laboratory parameters, including fasting blood glucose and blood lipids (total cholesterol, triglycerides, high-density lipoprotein, and low-density lipoprotein), were collected at baseline, week 4, and week 6.

[0034] According to the Chinese guidelines for the prevention and treatment of type 2 diabetes, individuals meeting any three of the following five criteria are defined as having MetS: (1) waist circumference ≥90 cm for men or ≥85 cm for women; (2) fasting blood glucose ≥6.10 mmol / L and / or use of hypoglycemic drugs; (3) blood pressure ≥130 / 85 mmHg and / or use of antihypertensive drugs; (4) triglycerides ≥1.70 mmol / L; (5) high-density lipoprotein cholesterol <1.04 mmol / L. In this embodiment, individuals who already met the baseline criteria for MetS were excluded. Newly diagnosed MetS within 6 weeks of follow-up was defined as drug-induced MetS for use in genetic and proteomics analysis.

[0035] The validation cohort was another prospective, multicenter, 6-week follow-up study that included patients aged 18-45 years with schizophrenia (including first-episode and chronic patients with acute exacerbations). The diagnostic criteria and severity requirements were consistent with the discovery cohort. All subjects gradually discontinued their original antipsychotic medications within one week of enrollment and uniformly switched to oral paliperidone extended-release formulation, with the dose adjusted within the range of 3-12 mg / day according to clinical condition over two weeks until it stabilized and remained unchanged. Clinical information and metabolic indicators were collected in the same manner as in the discovery cohort. Due to the limited number of MetS occurrences in this cohort, this embodiment used the percentage change in weight, waist circumference, blood pressure, blood glucose, and blood lipids relative to baseline over 6 weeks as continuous outcomes to evaluate and validate the performance of the multi-omics risk prediction model.

[0036] S120. Obtain genomic data from the discovery cohort and validation cohort, and combine them with clinical information and metabolic index data to perform genome-wide association analysis on the discovery cohort, screen for significant SNP sites that are significantly associated with antipsychotic drug-induced metabolic syndrome, and obtain polygenic risk score data for the validation cohort based on this.

[0037] In one embodiment, S120 may include: Preprocess the genomic data; Based on the genomic data of the discovery cohort, the genetic principal component data of the discovery cohort were obtained; Using the discovery of new-onset antipsychotic drug-induced metabolic syndrome within a set treatment period as the endpoint, clinical data and principal component data of the discovery cohort were used as covariates to construct a logistic regression model, and genome-wide association analysis was performed to screen out significant SNP sites that are significantly associated with antipsychotic drug-induced metabolic syndrome. Based on the results of genome-wide association analysis, significant SNP sites were located to genes, and a multigene risk score was constructed in the validation cohort. Sets of significant SNP sites under different P-value thresholds were screened, and the set of significant SNP sites under the optimal P-value threshold was obtained through fitting. The multigene risk score data of the validation cohort under the set of significant SNP sites corresponding to the optimal P-value threshold were obtained.

[0038] In this embodiment, the discovery cohort used Illumina Human Omni ZhongHua-8 Beadchips for genotyping. Standardized quality control was performed on the raw genotype data: at the sample level, samples with a genotype detection rate <98%, discrepancies between sex identification and registration, presence of first- or second-degree kinship, or identified as genetic anomalies were removed; at the locus level, SNPs with a minor allele frequency (MAF) <0.01, a detection rate <98%, or a Hardy–Weinberg equilibrium test p <1×10^-5 were removed. Subsequently, phase inference was performed using SHAPEIT, and genotype imputation was performed using IMPUTE2 combined with the 1000 Genomes Project Phase 1 (Version 3) population haplotypes, removing loci with an info score <0.9.

[0039] In the discovery cohort, with the endpoint of newly diagnosed APs-induced MetS within 6 weeks, genome-wide association analysis (GWAS) was performed using PLINK v1.9. A logistic regression model was constructed, with the top five principal components—sex, age, disease duration, type of antipsychotic medication, equivalent dose of chlorpromazine, research center, and population structure—as covariates to screen for SNPs significantly associated with APs-induced MetS. Based on the GWAS results of the male sample, this embodiment further constructed an APs-induced MetS-PRS, setting a series of p-value thresholds from 0 to 0.5 with a step size of 5 × 10^-5. Significant SNP sets under different thresholds were screened, and individual PRS were calculated. The threshold with the best fit was selected as the final model parameter. After standardization in the male sample of the validation cohort, the obtained PRS was used to test its predictive power for metabolic indicators such as weight and waist circumference percentage changes using linear regression, and the explanatory power of the model under different thresholds was compared.

[0040] S130. Obtain proteomic data from the validation cohort, locate significant SNP sites to genes to obtain candidate proteins, and combine the protein level data of the candidate proteins with their metabolic index data to perform metabolic outcome association analysis to obtain the protein level data of the target proteins associated with antipsychotic drug-induced metabolic syndrome in the validation cohort.

[0041] In one embodiment, S130 may include: Preprocess the proteomic data; Significant SNP sites are located to genes to obtain candidate proteins; Using the percentage change of metabolic parameters relative to baseline in the validation cohort over a set treatment period as the outcome, and clinical information from the validation cohort as covariates, a multiple linear regression model was constructed to conduct an association analysis of metabolic outcomes. This identified target proteins associated with the risk of antipsychotic drug-induced metabolic syndrome and obtained protein level data of target proteins associated with antipsychotic drug-induced metabolic syndrome in the validation cohort.

[0042] In this embodiment, baseline serum samples were collected from the validation cohort, and proteomic data were acquired using high-throughput mass spectrometry based on high-abundance protein removal and data-independent acquisition. First, a multi-affinity removal column (Top14 Abundant Protein Depletion) was used to remove 14 high-abundance proteins from the serum. After quantification using the BCA method, a certain amount of protein was precipitated, reduced, alkylated, and digested with trypsin, followed by desalting using a C18 column. Subsequently, DIA-MS analysis was performed using an EASY-nLC 1200 liquid chromatography system coupled to a Q Exactive HF-X mass spectrometer under 90-minute gradient elution conditions. Mass spectrometry data were analyzed in Spectronaut software and matched against the human UNIPROT database. Fixed modifications included cysteine ​​carbamide methylation, and variable modifications included N-terminal acetylation and methionine oxidation. The restriction enzyme sites were Trypsin / P, with a maximum of two missed cleavages allowed. Peptide length was limited to 7–52 amino acid residues, and the free-dip ratio (FDR) for peptide-spectrum matching and protein quantification was controlled to within 1%.

[0043] After quality control and filtering out proteins with low detection rates, only proteins with a detection rate exceeding 25% in all samples were retained for subsequent analysis. Missing values ​​were imputed using the median of the same protein in the remaining samples, and protein abundance was standardized. Based on this, a multiple linear regression model was constructed using percentage changes in body weight, waist circumference, and other metabolic indicators over 6 weeks as outcomes. Covariates such as age and disease duration were adjusted to systematically evaluate the association between baseline protein levels and APs-related metabolic adverse reactions, laying the foundation for screening candidate proteins with predictive value.

[0044] S140. Using the multigene risk score data of the validation cohort and / or the protein level data of the target protein of the validation cohort as independent variables, and the percentage change of the metabolic index data of the validation cohort relative to the baseline within a set treatment period as risk labels, a model for predicting the risk of metabolic syndrome caused by antipsychotic drugs is trained.

[0045] In one embodiment, S140 may include: The first prediction model was trained using the multigene risk score data of the validation cohort as the independent variable and the percentage change of the metabolic index data of the validation cohort relative to the baseline as the dependent variable. The second prediction model was trained by using the protein level data of the target protein in the validation cohort as the independent variable and the percentage change of the metabolic index data of the validation cohort relative to the baseline as the dependent variable. By combining multi-gene risk score data from the validation cohort and protein level data of the target protein as independent variables, and using the percentage change of metabolic index data from the validation cohort relative to the baseline as the risk label, a third prediction model was trained. The predictive performance of the first, second, and third prediction models was compared to obtain the preferred prediction model, which was then used as the final model for predicting the risk of metabolic syndrome caused by antipsychotic drugs.

[0046] The final model used to predict the risk of metabolic syndrome caused by antipsychotic drugs is the third prediction model.

[0047] This embodiment, after obtaining APs-induced MetS-PRS and related proteins, integrates these two types of information to construct a multi-omics risk prediction model. Specifically, it includes: (1) Construct a genetic model containing only PRS, with standardized PRS as the independent variable and percentage change of metabolic indicators as the dependent variable; (2) Construct a proteome model that only includes the abundance of candidate proteins, with standardized protein levels as the independent variable; (3) Construct a joint model, incorporating PRS and protein abundance into the same regression framework, adjusting for age, disease course and population structure principal components, and comparing the differences of the three models in terms of explanatory power (adjusted R²), root mean square error of residuals (RMSE) and goodness of fit.

[0048] By comparing the performance of the combined model with that of a single omics model, this embodiment demonstrates that integrating genomic and proteomic data can significantly improve the predictive ability of the risk of weight gain and waist circumference increase caused by antipsychotic drugs, thereby forming a multi-omics risk scoring method that can be promoted in clinical practice.

[0049] In this embodiment, after rigorous quality control and exclusion of participants from the baseline MetS, the cohort ultimately included 1,956 schizophrenia patients with complete genomic data (965 males and 991 females), and the validation cohort included 86 males and 114 females.

[0050] In this study, based on 4,609,700 autosomal SNPs, genome-wide association analysis (GWAS) was performed on APs-induced MetS in the whole sample, male samples, and female samples. The results showed a genome-wide significant risk locus in male patients: the intergenic SNP rs73762168 located at 6q21 (A1=T, MAF=0.03, P=1.77×10⁻⁸). Neighboring genes include SNX3, NR2E1, and AFG1L / LACE1. A potential association between rs73762168 and APs-induced MetS was also observed in the whole sample (P=8.46×10⁻⁵), while no significant association was observed in female samples, suggesting that this locus may have a male-specific genetic effect. Further region analysis based on linkage disequilibrium (LD) structure revealed that the high LD region containing rs73762168 simultaneously covers three genes: NR2E1, SNX3, and AFG1L / LACE1, all of which showed near-significant associations at the gene level. In the validation cohort, this example further investigated the impact of rs73762168 on actual metabolic outcomes. Results showed that male patients carrying the rs73762168_T allele exhibited more significant weight gain (P=6.06×10^-3) and waist circumference increase (P=9.22×10^-3) after 6 weeks of paliperidone treatment, exhibiting a heavier metabolic burden compared to patients with the GG genotype. However, the association between this locus and other metabolic indicators (such as blood pressure, blood glucose, and blood lipids) did not reach statistical significance. These results, based on follow-up data from an independent cohort, provide clinical evidence for the risk role of rs73762168 and its surrounding genes in APs-induced MetS.

[0051] In proteomics analysis, this embodiment used DIA-MS technology for high-throughput protein quantification in baseline serum samples of the validation cohort, and included proteins with a detection rate greater than 25% after strict quality control. Results showed that baseline SNX3 protein levels were significantly associated with changes in weight and waist circumference related to antipsychotic medications: after adjusting for covariates such as age and disease duration, higher SNX3 protein abundance was associated with a greater percentage increase in weight over 6 weeks (Beta=1.33, P=0.03) and a more significant percentage increase in waist circumference (Beta=1.29, P=8.87×10^-3). Although some associations did not survive under the most stringent multiple correction, their direction and effect were highly consistent with the SNP and gene-level analysis related to rs73762168, and statistically reached the conventional significance threshold, providing independent evidence at the protein level for SNX3 as a male-specific risk factor for APs-induced MetS. It is worth noting that the proteins of NR2E1 and AFG1L / LACE1 were not reliably detected in this serum proteomic analysis, therefore their protein levels in the circulatory system cannot be directly evaluated.

[0052] Building upon this, this embodiment constructs a PRS targeting the APs-induced MetS phenotype and focuses on evaluating its predictive ability in a male sample from the validation cohort. Results show that individual PRS significantly predicts weight and waist circumference gain after 6 weeks of paliperidone treatment: the regression coefficient between PRS and percentage weight gain was Beta = 1.92 (P = 7.93 × 10⁻⁴), and the regression coefficient between PRS and percentage waist circumference gain was Beta = 1.47 (P = 2.36 × 10⁻³), suggesting that genetic susceptibility has a quantifiable contribution to antipsychotic drug-related metabolic adverse reactions.

[0053] To highlight the gains from multi-omics integration, this embodiment further incorporates both PRS and baseline SNX3 protein levels into the prediction model, comparing the performance of single-omics and combined models. In the model containing only PRS, the explained value for weight gain was adjusted R² = 0.13; the explained value for the model containing only SNX3 protein was R² = 0.07. However, after incorporating both PRS and SNX3 protein into the same linear regression model, the combined model's explained value for weight gain increased to adjusted R² = 0.18, and the model fit was significantly better than the single PRS model (combined model vs. PRS model, P = 0.02) and the single protein model (combined model vs. protein model, P = 8.24 × 10⁻⁴). Regarding prediction error, the root mean square error (RMSE = 4.80) of the combined model was lower than that of the PRS model (4.99) and the protein model (5.17), further demonstrating that integrating genomic and proteomic information can substantially improve the prediction accuracy of antipsychotic drug-induced weight gain.

[0054] Based on a large sample of schizophrenia cases in China, this invention proposes and validates a multi-omics risk prediction method that integrates genomic and proteomic information, enabling quantitative assessment of the risk of metabolic syndrome induced by antipsychotic drugs. Genomic PRS can capture an individual's genetic susceptibility to adverse metabolic reactions; proteomic analysis based on baseline plasma protein levels reflects an individual's functional status at the metabolic and signaling pathway levels. By jointly modeling both, this invention significantly improves the predictive accuracy of phenotypes such as antipsychotic drug-related weight gain and waist circumference increase, providing a feasible technical approach for identifying high-risk patients before or early in the course of medication.

[0055] Clinically, the method of this invention can be used to: perform genotyping and a single serum protein test on patients before initiating antipsychotic drug treatment to predict individualized multi-omics risk scores; based on this, it can help clinicians select drug regimens with lower metabolic risks or more conservative dosage adjustment strategies, and strengthen the monitoring and early intervention of weight, waist circumference, and metabolic indicators for high-risk patients, thereby minimizing the risk of MetS while ensuring the efficacy of antipsychotic treatment, and improving long-term patient adherence and quality of life. In the future, this method can also be integrated with electronic medical record systems to form a semi-automated prediction and early warning tool, providing important support for precision medication and individualized follow-up in psychiatry.

[0056] The following describes the risk prediction system for metabolic syndrome caused by antipsychotic drugs provided by the present invention. The risk prediction system for metabolic syndrome caused by antipsychotic drugs described below can be referred to in correspondence with the construction method of the model for predicting the risk of metabolic syndrome caused by antipsychotic drugs described above.

[0057] See Figure 2 The present invention provides a risk prediction system for metabolic syndrome caused by antipsychotic drugs, which may include: The test data acquisition module is used to receive clinical data, metabolic index data, genomic data and proteomic data of the test subject from at least one terminal, wherein the test subject is a schizophrenic patient who is preparing to receive a certain antipsychotic drug treatment plan; The data processing module is used to: according to the method for constructing a model for predicting the risk of metabolic syndrome caused by antipsychotic drugs as described in any of the above, obtain the multigene risk score data of the subject under the set of significant SNP sites corresponding to the optimal P-value threshold based on the subject's genomic data; The prediction module is used to: based on the multigene risk score data of the subject under the set of significant SNP sites corresponding to the optimal P-value threshold, as well as clinical data, metabolic index data, and protein data, and through a model used to predict the risk of metabolic syndrome caused by antipsychotic drugs, obtain the prediction result of the subject's risk of metabolic syndrome caused by a certain antipsychotic drug treatment regimen. The prediction result output module is used to output the prediction results of the risk of metabolic syndrome caused by antipsychotic drugs under a certain antipsychotic drug treatment regimen to at least one terminal.

[0058] Figure 3 An example is a schematic diagram of the physical structure of an electronic device, such as... Figure 3 As shown, the electronic device may include a processor 810, a communications interface 820, a memory 830, and a communication bus 840, wherein the processor 810, the communications interface 820, and the memory 830 communicate with each other via the communication bus 840. The processor 810 can call logical instructions in the memory 830 to execute the following steps: The test subjects are schizophrenic patients who are preparing to receive a certain antipsychotic drug treatment regimen. According to the method for constructing a model for predicting the risk of metabolic syndrome caused by antipsychotic drugs as described above, based on the genomic data of the subject, the multigene risk score data of the subject under the set of significant SNP sites corresponding to the optimal P-value threshold is obtained. Based on the multigene risk score data of the subjects under the set of significant SNP sites corresponding to the optimal P-value threshold, as well as clinical data, metabolic index data, and protein data, the risk prediction results of metabolic syndrome caused by antipsychotic drugs under a certain antipsychotic drug treatment regimen are obtained by using a model for predicting the risk of metabolic syndrome caused by antipsychotic drugs.

[0059] Furthermore, the logical instructions in the aforementioned memory 830 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0060] On the other hand, the present invention also provides a computer program product, the computer program product comprising a computer program, the computer program being able to be stored on a non-transitory computer-readable storage medium, and the computer program being executed by a processor, enabling the computer to perform the following steps: The test subjects are schizophrenic patients who are preparing to receive a certain antipsychotic drug treatment regimen. According to the method for constructing a model for predicting the risk of metabolic syndrome caused by antipsychotic drugs as described above, based on the genomic data of the subject, the multigene risk score data of the subject under the set of significant SNP sites corresponding to the optimal P-value threshold is obtained. Based on the multigene risk score data of the subjects under the set of significant SNP sites corresponding to the optimal P-value threshold, as well as clinical data, metabolic index data, and protein data, the risk prediction results of metabolic syndrome caused by antipsychotic drugs under a certain antipsychotic drug treatment regimen are obtained by using a model for predicting the risk of metabolic syndrome caused by antipsychotic drugs.

[0061] In another aspect, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, performs the following steps: The test subjects are schizophrenic patients who are preparing to receive a certain antipsychotic drug treatment regimen. According to the method for constructing a model for predicting the risk of metabolic syndrome caused by antipsychotic drugs as described above, based on the genomic data of the subject, the multigene risk score data of the subject under the set of significant SNP sites corresponding to the optimal P-value threshold is obtained. Based on the multigene risk score data of the subjects under the set of significant SNP sites corresponding to the optimal P-value threshold, as well as clinical data, metabolic index data, and protein data, the risk prediction results of metabolic syndrome caused by antipsychotic drugs under a certain antipsychotic drug treatment regimen are obtained by using a model for predicting the risk of metabolic syndrome caused by antipsychotic drugs.

[0062] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0063] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0064] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for constructing a model for predicting the risk of metabolic syndrome induced by antipsychotic drugs, characterized in that, include: Clinical information and metabolic data were obtained from the discovery cohort and the validation cohort, respectively. The discovery cohort consisted of schizophrenia patients receiving different antipsychotic drug treatments, while the validation cohort consisted of schizophrenia patients receiving uniform antipsychotic drug treatments. Both the discovery cohort and the validation cohort included patients who developed new-onset antipsychotic drug-induced metabolic syndrome within a set treatment period. Genomic data from the discovery and validation cohorts were obtained, and combined with their clinical information and metabolic marker data, genome-wide association analysis was performed on the discovery cohort to screen for significant SNP sites that are significantly associated with antipsychotic drug-induced metabolic syndrome. Based on this, polygenic risk score data for the validation cohort were obtained. Proteomic data of the validation cohort were obtained, significant SNP sites were located to genes to obtain candidate proteins, and metabolic outcome association analysis was performed by combining the protein level data of the candidate proteins with their metabolic index data to obtain the protein level data of the target proteins associated with antipsychotic drug-induced metabolic syndrome in the validation cohort. Using multigene risk score data from the validation cohort and / or protein level data of the target protein from the validation cohort as independent variables, and the percentage change of metabolic indicators from the baseline within a set treatment period from the validation cohort as risk labels, a model for predicting the risk of metabolic syndrome induced by antipsychotic drugs was trained.

2. The method for constructing a model for predicting the risk of metabolic syndrome induced by antipsychotic drugs according to claim 1, characterized in that, Clinical information includes any one or any combination of the following: demographic information, height and weight, waist circumference, type of antipsychotic medication, and dosage of antipsychotic medication; Metabolic indicators include fasting blood glucose and / or blood lipids.

3. The method for constructing a model for predicting the risk of metabolic syndrome induced by antipsychotic drugs according to claim 2, characterized in that, The process involves acquiring genomic data from the discovery and validation cohorts, combining this data with clinical information and metabolic marker data, performing genome-wide association analysis on the discovery cohort, screening for significant SNPs associated with antipsychotic drug-induced metabolic syndrome, and obtaining polygenic risk score data for the validation cohort based on this. Preprocess the genomic data; Based on the genomic data of the discovery cohort, the genetic principal component data of the discovery cohort were obtained; Using the discovery of new-onset antipsychotic drug-induced metabolic syndrome within a set treatment period as the endpoint, clinical data and principal component data of the discovery cohort were used as covariates to construct a logistic regression model, and genome-wide association analysis was performed to screen out significant SNP sites that are significantly associated with antipsychotic drug-induced metabolic syndrome. Based on the results of genome-wide association analysis, significant SNP sites were located to genes, and a multigene risk score was constructed in the validation cohort. Sets of significant SNP sites under different P-value thresholds were screened, and the set of significant SNP sites under the optimal P-value threshold was obtained through fitting. The multigene risk score data of the validation cohort under the set of significant SNP sites corresponding to the optimal P-value threshold were obtained.

4. The method for constructing a model for predicting the risk of metabolic syndrome induced by antipsychotic drugs according to claim 3, characterized in that, The set of significant SNP sites includes the intergenic SNP rs73762168 located at 6q21.

5. The method for constructing a model for predicting the risk of metabolic syndrome induced by antipsychotic drugs according to claim 3, characterized in that, The process involves acquiring proteomic data from the validation cohort, locating significant SNP sites to genes to obtain candidate proteins, and combining the protein level data of the candidate proteins with their metabolic index data to perform metabolic outcome association analysis. This yields protein level data of target proteins in the validation cohort associated with antipsychotic drug-induced metabolic syndrome, including: Preprocess the proteomic data; Significant SNP sites are located to genes to obtain candidate proteins; Using the percentage change of metabolic parameters relative to baseline in the validation cohort over a set treatment period as the outcome, and clinical information from the validation cohort as covariates, a multiple linear regression model was constructed to conduct an association analysis of metabolic outcomes. This identified target proteins associated with the risk of antipsychotic drug-induced metabolic syndrome and obtained protein level data of target proteins associated with antipsychotic drug-induced metabolic syndrome in the validation cohort.

6. The method for constructing a model for predicting the risk of metabolic syndrome induced by antipsychotic drugs according to claim 5, characterized in that, The target protein is SNX3.

7. The method for constructing a model for predicting the risk of metabolic syndrome induced by antipsychotic drugs according to claim 5, characterized in that, The model, trained using multigene risk score data from the validation cohort and / or protein level data of the target protein from the validation cohort as independent variables, and the percentage change of metabolic indicators from baseline within a set treatment period from the validation cohort as a risk label, for predicting the risk of metabolic syndrome induced by antipsychotic drugs, includes: The first prediction model was trained using the multigene risk score data of the validation cohort as the independent variable and the percentage change of the metabolic index data of the validation cohort relative to the baseline as the dependent variable. The second prediction model was trained by using the protein level data of the target protein in the validation cohort as the independent variable and the percentage change of the metabolic index data of the validation cohort relative to the baseline as the dependent variable. By combining multi-gene risk score data from the validation cohort and protein level data of the target protein as independent variables, and using the percentage change of metabolic index data from the validation cohort relative to the baseline as the risk label, a third prediction model was trained. The predictive performance of the first, second, and third prediction models was compared to obtain the preferred prediction model, which was then used as the final model for predicting the risk of metabolic syndrome caused by antipsychotic drugs.

8. The method for constructing a model for predicting the risk of metabolic syndrome induced by antipsychotic drugs according to claim 7, characterized in that, The final model used to predict the risk of metabolic syndrome caused by antipsychotic drugs is the third prediction model.

9. A risk prediction system for metabolic syndrome induced by antipsychotic drugs, characterized in that, include: The test data acquisition module is used to receive clinical data, metabolic index data, genomic data and proteomic data of the test subject from at least one terminal, wherein the test subject is a schizophrenic patient who is preparing to receive a certain antipsychotic drug treatment plan; The data processing module is used to: construct a model for predicting the risk of metabolic syndrome caused by antipsychotic drugs according to any one of claims 1-8, and obtain multigene risk score data of the subject under the set of significant SNP sites corresponding to the optimal P-value threshold based on the subject's genomic data; The prediction module is used to: based on the multigene risk score data of the subject under the set of significant SNP sites corresponding to the optimal P-value threshold, as well as clinical data, metabolic index data, and protein data, and through a model used to predict the risk of metabolic syndrome caused by antipsychotic drugs, obtain the prediction result of the subject's risk of metabolic syndrome caused by a certain antipsychotic drug treatment regimen. The prediction result output module is used to output the prediction results of the risk of metabolic syndrome caused by antipsychotic drugs under a certain antipsychotic drug treatment regimen to at least one terminal.

10. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the method for constructing a model for predicting the risk of metabolic syndrome induced by antipsychotic drugs as described in any one of claims 1 to 8, and / or, the following steps: The test subjects are schizophrenic patients who are preparing to receive a certain antipsychotic drug treatment regimen. The method for constructing a model for predicting the risk of metabolic syndrome caused by antipsychotic drugs according to any one of claims 1-8, obtains multigene risk score data of the subject under the set of significant SNP sites corresponding to the optimal P-value threshold based on the subject's genomic data. Based on the multigene risk score data of the subjects under the set of significant SNP sites corresponding to the optimal P-value threshold, as well as clinical data, metabolic index data, and protein data, the risk prediction results of metabolic syndrome caused by antipsychotic drugs under a certain antipsychotic drug treatment regimen are obtained by using a model for predicting the risk of metabolic syndrome caused by antipsychotic drugs.

11. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the method for constructing a model for predicting the risk of metabolic syndrome induced by antipsychotic drugs as described in any one of claims 1 to 8, and / or, the following steps: The test subjects are schizophrenic patients who are preparing to receive a certain antipsychotic drug treatment regimen. The method for constructing a model for predicting the risk of metabolic syndrome caused by antipsychotic drugs according to any one of claims 1-8, obtains multigene risk score data of the subject under the set of significant SNP sites corresponding to the optimal P-value threshold based on the subject's genomic data. Based on the multigene risk score data of the subjects under the set of significant SNP sites corresponding to the optimal P-value threshold, as well as clinical data, metabolic index data, and protein data, the risk prediction results of metabolic syndrome caused by antipsychotic drugs under a certain antipsychotic drug treatment regimen are obtained by using a model for predicting the risk of metabolic syndrome caused by antipsychotic drugs.

Images