Mendel randomization-based mental disease drug targeted analysis method and system

Abstract

The invention discloses a Mendel randomization-based mental disease drug targeted analysis method and system, and the method comprises the steps: carrying out the analysis processing through a Mendel randomization analysis process, and obtaining the causal effect estimation of each target gene corresponding to a plurality of phenotypes; performing multiple inspection correction on the causal effect estimation of each target gene corresponding to multiple phenotypes to obtain a significant target gene after correction; carrying out sensitivity analysis on the corrected significant target genes to obtain the target genes screened for the first time; performing co-localization analysis on the target genes screened for the first time to obtain target genes screened for the second time; carrying out pathway enrichment analysis on the target genes screened for the second time, then carrying out protein interaction network analysis, and finally carrying out drug matching on the target genes of the protein interaction network to obtain an old drug new use prediction result or taking the target genes as the target genes for guiding new drug preparation.

Description

Technical Field

[0001] The present invention relates to the technical field of drug targeting analysis, and in particular to a method and system for targeting analysis of drugs for mental illness based on Mendelian randomization. Background Art

[0002] The high incidence of mental illness not only has a serious impact on the lives of individual patients, but also brings a heavy burden to families and society. The complexity of mental illness comes from the diversity of its causes and the uncertainty of the course of the disease, including the combined effects of multiple factors such as genetics, environment, lifestyle, and psychological factors. The causes of mental illness often involve multiple biological pathways and psychosocial factors, such as gene mutations, neurotransmitter imbalances, abnormal brain structure and function, etc. The uncertainty of the course of the disease is reflected in the volatility and recurrence of the disease and the individual differences in response to treatment. Although drug therapy is one of the main means of treating mental illness, the efficacy of existing drugs is often limited and accompanied by varying degrees of side effects.

[0003] There are many challenges in evaluating the efficacy and safety of drugs, including the unclear causal relationship between drugs and diseases, the complexity of drug mechanisms of action, and the uncertainty of drug-disease interactions. Mental illness not only affects the quality of life of patients, but also leads to reduced work ability and impaired social function, thereby increasing health care costs and social welfare burdens.

[0004] The problems existing in the prior art are:

[0005] 1. The ambiguity of the causal relationship between drugs and diseases: Traditional drug research methods often have difficulty revealing the exact causal relationship between drugs and diseases, which limits the precise selection of drug targets and the formulation of treatment strategies. For example, some drugs may improve symptoms in the short term, but long-term use may cause new health problems.

[0006] 2. Complexity of data integration and analysis: Multi-omics data of mental illness, such as genome, transcriptome, proteome and metabolome data, are highly complex and heterogeneous. The lack of effective data integration and analysis tools makes it difficult to extract useful information from these data, thus affecting the efficiency and accuracy of drug development. Summary of the invention

[0007] In order to address the deficiencies of the prior art, the present invention provides a method and system for targeted analysis of psychiatric drugs based on Mendelian randomization; the invention aims to construct an analysis platform integrating multi-omics data through a drug-targeted Mendelian randomization method to solve the problems of unclear causal relationship between drugs and diseases, complex data integration and analysis, and lack of personalized treatment plans.

[0008] On the one hand, a method for targeted analysis of drugs for psychiatric disorders based on Mendelian randomization is provided, comprising:

[0009] S101: obtaining an existing proteomic dataset and an existing transcriptomic dataset; performing preprocessing operations on the existing proteomic dataset and the existing transcriptomic dataset to obtain a preprocessed proteomic dataset and a preprocessed transcriptomic dataset; obtaining a target mental illness dataset, an existing drug database, and an existing phenotype database;

[0010] S102: Based on the target mental illness dataset, existing drug database, existing phenotype database, pre-processed proteome dataset and pre-processed transcriptome dataset, the Mendelian randomization analysis process is used to analyze and process, and the causal effect estimation of each target gene corresponding to multiple phenotypes is obtained;

[0011] S103: For each target gene, the causal effect of multiple phenotypes is estimated, and multiple test correction is performed to obtain the target genes that are significant after correction; for the target genes that are significant after correction, sensitivity analysis is performed to obtain the target genes for the first screening; for the target genes for the first screening, co-localization analysis is performed to obtain the target genes for the second screening;

[0012] S104: For the target genes screened for the second time, pathway enrichment analysis is first performed, followed by protein interaction network analysis, and finally drug matching is performed on the target genes of the protein interaction network to obtain the prediction results of old drug new use or use the target genes as target genes for guiding the preparation of new drugs. Old drug new use means using a drug already on the market to treat the target mental illness.

[0013] On the other hand, a system for drug targeting analysis of mental illness based on Mendelian randomization is provided, including:

[0014] An acquisition module is configured to: acquire an existing proteomic dataset and an existing transcriptomic dataset; perform preprocessing operations on the existing proteomic dataset and the existing transcriptomic dataset to obtain a preprocessed proteomic dataset and a preprocessed transcriptomic dataset; acquire a target mental illness dataset, an existing drug database, and an existing phenotype database;

[0015] An analysis module is configured to: perform analysis and processing using a Mendelian randomization analysis process based on a target mental illness dataset, an existing drug database, an existing phenotype database, a preprocessed proteome dataset, and a preprocessed transcriptome dataset to obtain causal effect estimates of each target gene corresponding to multiple phenotypes;

[0016] The screening module is configured as follows: for each target gene, causal effect estimation corresponding to multiple phenotypes is performed, multiple test correction is performed, and the target genes that are significant after correction are obtained; for the target genes that are significant after correction, sensitivity analysis is performed to obtain the target genes for the first screening; for the target genes for the first screening, co-localization analysis is performed to obtain the target genes for the second screening;

[0017] The output module is configured as follows: for the target genes screened for the second time, pathway enrichment analysis is first performed, followed by protein interaction network analysis, and finally drug matching is performed on the target genes of the protein interaction network to obtain the prediction results of old drug new use or use the target genes as target genes for guiding the preparation of new drugs. Old drug new use means using drugs already available on the market to treat target mental illnesses.

[0018] On the other hand, there is also provided an electronic device, comprising:

[0019] a memory for non-transitory storage of computer-readable instructions; and

[0020] a processor for executing the computer readable instructions,

[0021] When the computer-readable instructions are executed by the processor, the method described in the first aspect is executed.

[0022] On the other hand, a storage medium is provided, which non-temporarily stores computer-readable instructions, wherein when the non-temporary computer-readable instructions are executed by a computer, the instructions of the method described in the first aspect are executed.

[0023] On the other hand, a computer program product is provided, comprising a computer program, wherein the computer program is used to implement the method described in the first aspect when running on one or more processors.

[0024] The above technical solution has the following advantages or beneficial effects:

[0025] 1. Evaluation of drug efficacy and safety: Through the Mendelian randomization method, using genetic variation as a natural experimental tool, the causal relationship between drugs and diseases can be more accurately evaluated, thus providing a scientific basis for the selection of drug targets and the formulation of treatment strategies.

[0026] 2. Data integration and analysis: This platform will integrate multi-omics data and use advanced bioinformatics and statistical methods to achieve effective data integration and analysis. This will help reveal the molecular mechanisms of mental illness and provide new perspectives for drug development.

[0027] 3. Improve treatment efficiency: By optimizing the drug screening and evaluation process, this platform can shorten the drug development cycle and speed up the launch of new drugs, thereby providing patients with effective treatment options more quickly.

[0028] 4. Reduce medical costs: By improving the efficacy and safety of drugs and reducing unnecessary medical interventions and side effects of treatment, this platform helps to reduce overall medical costs and alleviate the financial burden on patients.

[0029] 5. Promote interdisciplinary collaboration: The development and application of this platform will promote collaboration among experts in multiple disciplines such as medicine, genetics, bioinformatics, and pharmacology, and jointly promote progress in the treatment of mental illness and drug development. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] The accompanying drawings in the specification, which constitute a part of the present invention, are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute improper limitations on the present invention.

[0031] Figure 1 is a flow chart of the method of embodiment 1;

[0032] Figure 2 The Mendelian randomization causal diagram model of Example 1;

[0033] Figure 3 A plan for building a drug targeting platform in Example 1;

[0034] Figure 4 The process of building the drug targeting platform of Example 1;

[0035] Figure 5 The drug-targeted MR screening tool variable process of Example 1;

[0036] Figure 6 This is a schematic diagram of the results of using the drug-targeted Mendelian randomization platform data in Example 1. DETAILED DESCRIPTION

[0037] It should be noted that the following detailed descriptions are exemplary and are intended to provide further explanation of the present invention. Unless otherwise specified, all technical and scientific terms used herein have the same meanings as those commonly understood by those skilled in the art to which the present invention belongs.

[0038] In the absence of conflict, the embodiments of the present invention and the features of the embodiments may be combined with each other.

[0039] In this embodiment, all data is obtained in compliance with laws and regulations and based on the user's consent, and is used legally.

[0040] Embodiment 1

[0041] like Figure 1 As shown, this embodiment provides a targeted analysis method for psychiatric drug based on Mendelian randomization;

[0042] Targeted analysis of psychiatric drug treatments based on Mendelian randomization, including:

[0043] S101: obtaining an existing proteomic dataset and an existing transcriptomic dataset; performing preprocessing operations on the existing proteomic dataset and the existing transcriptomic dataset to obtain a preprocessed proteomic dataset and a preprocessed transcriptomic dataset; obtaining a target mental illness dataset, an existing drug database, and an existing phenotype database;

[0044] S102: Based on the target mental illness dataset, existing drug database, existing phenotype database, pre-processed proteome dataset and pre-processed transcriptome dataset, the Mendelian randomization analysis process is used to analyze and process, and the causal effect estimation of each target gene corresponding to multiple phenotypes is obtained;

[0045] S103: For each target gene, the causal effect of multiple phenotypes is estimated, and multiple test correction is performed to obtain the target genes that are significant after correction; for the target genes that are significant after correction, sensitivity analysis is performed to obtain the target genes for the first screening; for the target genes for the first screening, co-localization analysis is performed to obtain the target genes for the second screening;

[0046] S104: For the target genes screened for the second time, pathway enrichment analysis is first performed, followed by protein interaction network analysis. Finally, drug matching is performed on the target genes in the protein interaction network to obtain the results of new uses of old drugs or target genes used to guide the preparation of new drugs.

[0047] Furthermore, the method further comprises:

[0048] S105: Conduct safety evaluation on target genes used to guide the preparation of new drugs and assess the safety of target genes.

[0049] Furthermore, the S101: obtaining an existing proteomic dataset and an existing transcriptomic dataset; wherein the proteomic dataset refers to: proteomic pQTL data; wherein the transcriptomic dataset refers to: transcriptomic eQTL data.

[0050] It should be understood that the proteomics pQTL data refers to: The UK Biobank Pharma Preteomics Project (UKB-PPP) is a competitive alliance of 13 international biopharmaceutical companies that aims to generate high-dimensional, population-scale proteomics data. In this way, proteomics can be expanded to a large scale, expand the research use of high-throughput proteomics data, establish a broader pQTL database, and accelerate the discovery of biomarkers, diagnostic methods and drugs. The researchers used the antibody-based Olink Explore 3072PEA technology to measure, process and downstream analyze 2941 plasma analytes from 54,219 UK Biobank participants, and finally obtained pQTL positioning of 2923 proteins that could identify 14,287 major genetic associations. These findings show that proteomics has important value and significance in elucidating the biological mechanisms of diseases, exploring potential therapeutic targets, and promoting drug development. Large-scale protein genomics studies can increase the power and availability of genetic tools for Mendelian randomization, simulating the drug targeting effects observed in clinical trials [Sun BB, Chiou J, Traylor M, et al. Plasma proteomic associations with genetics and health in the UK Biobank. Nature. 2023 Oct; 622(7982): 329-338]. Therefore, the present invention uses a large-scale genome-wide protein database containing 2923 proteins in UKB-PPP as exposure data.

[0051] The proteomics component of the Atherosclerosis Risk in Communities (ARIC) database is a multidisciplinary biobank project dedicated to creating a large-scale proteome database related to cardiovascular disease. The ARIC project itself is a long-term, multicenter cohort study that initially included 15,792 adults who were recruited into the project between 1987 and 1989 and were aged 45-64. Using cutting-edge mass spectrometry technology, ARIC researchers conducted in-depth proteomic analysis of blood samples from thousands of participants, aiming to reveal the molecular basis of cardiovascular disease and promote the development of early diagnosis and treatment strategies. In this study, a high-resolution mass spectrometry platform was used to perform comprehensive protein expression profiling of blood samples in the ARIC cohort. The researchers successfully identified and quantified thousands of proteins, and through advanced bioinformatics tools, established a pQTL (protein quality trait locus) database containing rich genetic information. These data not only enhance our understanding of the molecular mechanisms of cardiovascular disease, but also provide valuable resources for the discovery of future biomarkers and the identification of drug targets. The proteomics section of the ARIC database particularly emphasizes the importance of large-scale biomarker discovery, and its data have been used in multiple genetic and functional studies to explore potential therapeutic targets for cardiovascular disease. In addition, the database provides powerful genetic tools for Mendelian randomization studies, which helps to simulate the effects of drug treatments in clinical trials, thereby evaluating the potential effects of drugs on cardiovascular disease without the risks of traditional interventional clinical trials [Wright, JD, Folsom, AR, Coresh, J., Sharrett, AR, Couper, D., Wagenknecht, LE, Mosley, TH, Jr, Ballantyne, CM, Boerwinkle, EA, Rosamond, WD, & Heiss, G. (2021). The ARIC (Atherosclerosis Risk In Communities) Study: JACC Focus Seminar 3 / 8. Journal of the American College of Cardiology, 77(23), 2939–2959].

[0052] It should be understood that the transcriptomics eQTL data refers to: eQTLGen Consortium is an international scientific cooperation organization jointly coordinated by the University of Groningen Medical Center, the University of Tartu, the Estonian Genome Center and other institutions, dedicated to in-depth research on the genetic structure of blood gene expression and exploring the genetic basis of complex traits. The consortium reveals the connection between gene expression and genetic variation through large-scale genome-wide association analysis. The eQTLGen database is a large eQTL database obtained by the consortium based on blood gene expression and meta-analysis at the overall gene level, including eQTL data of 19,250 genes from 31,684 participants in 37 independent cohort sample sets. Researchers have completed cis-eQTL, targeted trans-eQTL and eQTS analysis, testing the association of approximately 10,000 trait-related variants with distal gene expression. These research results have been published and widely cited, and the data are free and open. The database was established to promote collaboration and data coordination between different blood eQTL cohorts, use the high statistical power of the consortium to identify genes with weak genetic effects on blood gene expression, and provide a valuable resource for the scientific community. In addition, researchers can use these data to identify new biomarkers, develop diagnostic methods, and accelerate drug discovery. Large-scale gene expression genomics studies can enhance the power and usability of genetic tools for Mendelian randomization, mimicking drug target effects observed in clinical trials.

[0053] Genotype-Tissue Expression (GTEx) is a large-scale public resource project funded by the National Institutes of Health (NIH) of the United States. It provides data for researchers studying tissue- and cell-specific gene expression and regulation in individuals, development, and species. The data comes from three NIH projects: Adult GTEx, dGTEx, and Non-Human Primate Development GTEx (NHP-dGTEx). The Adult Genotype-Tissue Expression (GTEx) project we are currently using is a comprehensive public resource for studying human gene expression, regulation, and its relationship with genetic variation in a variety of different tissues and individuals. It aims to study the relationship between human genome variation and tissue-specific gene expression. The latest eighth version (V8) of the GTEx database was published in the journal Science in 2020. It includes DNA data from 838 healthy human donors (including Whole Genome Sequencing, WGS and Whole Exome Sequencing, WES) and 17,382 RNA-seq data from nearly 1,000 human individuals, covering 54 different tissue and organ sites, as well as 2 cell lines from donors' blood and skin. It is currently the most comprehensive database of healthy human tissue samples in the world. The project is of great value in understanding the genetic control mechanisms of gene expression, discovering new biomarkers, and studying the molecular basis of complex diseases. In addition, researchers can use the GTEx database to explore the expression patterns of specific genes in different tissues, identify changes in gene expression associated with specific diseases, and reveal the genetic basis of gene expression and the interaction between environmental factors. The present invention uses the gene transcription expression containing antihypertensive drug targets in the GTEx database as exposure data [GTEx Consortium. The GTEx Consortium atlas of genetic regulatory effects across human tissues. Science. 2020; 369(6509): 1318-1330. doi: 10.1126 / science.aaz1776].

[0054] Furthermore, the preprocessing operation is performed on the existing proteomic dataset and the existing transcriptomic dataset to obtain the preprocessed proteomic dataset and the preprocessed transcriptomic dataset, and the preprocessing operation specifically includes:

[0055] S101-1: Determine the significance of the whole genome P < 5e -8 Is it true? If so, the single nucleotide polymorphism candidate is used as an instrumental variable for subsequent processing; if not, the single nucleotide polymorphism is not used as an instrumental variable; P indicates genome-wide significance;

[0056] S101-2: Screening for cis-pQTLs from the single nucleotide polymorphisms obtained from S101-1; cis-pQTLs are defined as protein-coding genes within 1 Mb from 500 kb upstream of the start site to 500 kb downstream of the stop site, or protein-coding genes within 1 Mb from 500 kb upstream and downstream of the transcription start site;

[0057] S101-3: The single nucleotide polymorphisms screened by S101-2 were clustered by linkage disequilibrium (LD). The specific clustering parameter was r 2 = 0.01 The window is 1000 kb, and the 1000 Genomes Project Phase 3 (EUR) is used as the reference panel to ensure the independence of the selected SNPs: r 2 = 0.01, the window is 1Mb, that is, remove the r 2 SNPs with >0.01 are instrumental variables.

[0058] It should be understood that the exposure data were selected: according to the DrugBank, CheMBL, and DGIdb databases, the target genes of the target drugs were obtained. Subsequently, the proteomics databases ARIC and UKB-PPP, and the transcriptomics databases eqtlGEN and GTEx target genes were used as exposures;

[0059] Screening of effective instrumental variables: Screening of the exposure data obtained to meet the following criteria: (1) being located in the cis region; (2) being strongly correlated: the significance threshold is P < 5e -8 ; (3) Independence: Linkage disequilibrium (LD) clustering was performed using the 1000Genomes Project Phase 3 (EUR) as the reference panel to ensure the independence of the selected SNPs: r2<0.01, within1 Mb distance; (4) Removal of weak instrumental variables: The F statistic was used to quantify the strength of the instrumental variables, and instrumental variables with F>10 were selected.

[0060] Furthermore, the target mental illness dataset, existing drug database and existing phenotype database are obtained; wherein the target mental illness dataset includes: major depressive disorder, attention deficit hyperactivity disorder, post-traumatic stress disorder, schizophrenia, autism spectrum disorder, anorexia nervosa, bipolar disorder, neuroticism and anxiety.

[0061] Among them, there are existing drug databases, including: ChEMBL, DrugBank, and DGIdb databases;

[0062] It should be understood that after obtaining significant protein targets through MR analysis, further drug databases, including ChEMBL, DrugBank, and DGIdb databases, are queried to evaluate the drug development prospects of the targets identified in the present invention.

[0063] The ChEMBL database is a manually curated, high-quality, large-scale, open global core biological data resource (URL: https: / / www.ebi.ac.uk / chembl / ), providing comprehensive information about bioactive molecules with drug properties. Since its first release in 2009, ChEMBL has grown into a database containing diverse data types to support scientific research related to drug discovery, including academics, biotechnology companies and large global organizations. The data content of ChEMBL continues to grow, covering a wide range of aspects from the physicochemical properties of drugs and ligands, drug repurposing opportunities, identification of protein degradation targets, selectivity analysis of ligands, to machine learning-based target prediction models. In addition, ChEMBL also provides detailed information on approved drugs and clinical candidates, including their indications, mechanisms of action, black box warnings and withdrawal information. The ChEMBL database provides researchers with a powerful tool for drug discovery and chemical genomics research through its user interface and API. It is also an important resource for the development, validation and application of machine learning and artificial intelligence methods [Zdrazil B, Felix E, Hunter F, et al. The ChEMBL Database in 2023: a drug discovery platform spanning multiple bioactivity data types and timeperiods. Nucleic Acids Res. 2024 Jan 5; 52(D1): D1180-D1192].

[0064] The DrugBank database is a comprehensive and widely used knowledge resource that provides researchers and clinicians with detailed information about drugs, drug targets, and related drug interactions (URL: https: / / www.drugbank.ca / ). Since its first release in 2006, DrugBank has become the "gold standard" in the field of drug informatics and is widely used in a variety of biomedical research and clinical applications. DrugBank version 6.0 has significantly improved the amount and quality of data, including an increase in the number of FDA-approved drugs, investigational drugs, drug-drug interactions, and drug-food interactions. In addition, this version introduces new data types, such as detailed drug mechanisms of action and metabolic pathways, experimental and predicted mass spectrometry and nuclear magnetic resonance spectroscopy data, and improved user interfaces and search functions, making it an important tool for drug discovery, drug repurposing, understanding of drug mechanisms, and tracking of drug interactions. Through its publicly accessible database and web interface, DrugBank provides valuable data resources for academia, medicinal chemists and pharmaceutical companies, supporting the development of data-driven drug informatics and precision medicine [Knox C, Wilson M, Klinger CM, et al. DrugBank 6.0: the DrugBank Knowledgebase for 2024. Nucleic Acids Res. 2024 Jan 5; 52(D1): D1265-D1275.].

[0065] The Drug-Genne Interaction Database (DGIdb) is a public database designed to promote precision medicine and drug discovery. It provides a valuable resource for clinicians and researchers by aggregating gene or gene product, drug, and drug-gene interaction information (URL: http: / / dgidb.genome.wustl.edu / ). DGIdb 5.0 has undergone important updates, including the introduction of new data sources and improved user interfaces to support more efficient data access and queries. In addition, this version also uses a GraphQL API, allowing users to perform customized queries to more accurately obtain the required drug-gene interaction data. The architectural updates and functional enhancements of DGIdb have made it an indispensable tool in clinical research and drug development processes [Cannon M, Stevenson J, Stahl K, et al. DGIdb 5.0: rebuilding the drug-gene interaction database for precision medicine and drug discovery platforms. Nucleic Acids Res. 2024 Jan 5; 52(D1): D1227-D1235].

[0066] Among them, there are phenotype databases, including: phenotype SAIGE database.

[0067] SAIGE (Scalable and Accurate Implementation of Generalised Mixed Model, SAIGE) database, UK Biobank (UKB) is a resource containing biological samples and detailed information of 500,000 British participants, which provides valuable data for genetic research. In genome-wide association studies of thousands of binary phenotypes in large biological databases, the number of cases in most binary phenotypes is significantly less than that in the control group. When linear mixed models or logistic regression mixed models are used to analyze phenotypes with imbalanced case-control ratios, the model will produce a large type I error rate and easily cause false positives. Based on this problem, Zhou et al. proposed a scalable and accurate generalized mixed model, which has two major advantages: first, the saddlepoint approximation (SPA) is used to correct the distribution of the score test statistic, which can provide accurate P values ​​even when the case-control ratio is severely unbalanced; second, the most advanced optimization strategy is used to reduce the computational cost. Therefore, SAIGE is suitable for GWAS analysis of thousands of phenotypes in large biobanks, which can effectively solve the problem of imbalanced case-control ratio and sample correlation. It can effectively solve the problem of case-control imbalance and sample-relatedness [Efficiently controlling for case-control imbalance and sample-relatedness in large-scale genetic association studies]. The researchers used SAIGE to effectively analyze the genotype data of 1403 binary disease phenotypes and 28 million gene loci in the UKB, and obtained a large amount of summary statistics on disease traits, followed by genome-wide association study analysis, which helped to identify genetic variants associated with various diseases and phenotypes. The UKB SAIGE GWAS database is a powerful resource that supports genetic research worldwide by providing a wide range of genetic and phenotypic data, and has made important contributions to understanding the genetic basis of human disease.

[0068] Furthermore, the S102: based on the target mental illness dataset, the existing drug database, the existing phenotype database, the pre-processed proteome dataset and the pre-processed transcriptome dataset, the Mendelian randomization analysis process is used to perform analysis and processing to obtain the causal effect estimation of each target gene corresponding to multiple phenotypes, specifically including:

[0069] S102-1: From the GWAS-CATALOG database, GWAS data of mental illnesses were selected as outcome data, including major depressive disorder, attention deficit hyperactivity disorder, post-traumatic stress disorder, schizophrenia, autism spectrum disorder, anorexia nervosa, bipolar disorder, neuroticism, and anxiety;

[0070] S102-2: Two-sample Mendelian randomization analysis was performed using the R package TwoSampleMR to estimate the causal effect of each target gene on multiple psychiatric disorders.

[0071] The input files of two-sample Mendelian randomization analysis include: instrumental variables of target genes and outcomes of two-sample Mendelian randomization analysis;

[0072] The instrumental variables of the target genes are the SNPs screened out in steps S101-1 to S101-3, and the outcomes of the two-sample Mendelian randomization analysis are the GWAS data of mental illnesses selected in S102-1.

[0073] When there is only one instrumental variable, the Wald ratio method of Mendelian randomization is used to estimate the causal effect;

[0074] When there are two or more instrumental variables, the IVW (inverse-variance weighted) method of Mendelian randomization is used to estimate the causal effect.

[0075] It should be understood that the GWAS-CATALOG database [Sollis E, Mosaku A, Abid A, Buniello A, Cerezo M, Gil L, Groza T, O,Hall P,Hayhurst J,Ibrahim A,Ji Y,John S,LewisE,MacArthur JAL,McMahon A,Osumi-Sutherland D,Panoutsopoulou K,Pendlington Z,Ramachandran S,Stefancsik R,Stewart J,Whetzel P,Wilson R,Hindorff L,Cunningham F,Lambert SA,Inouye M,Parkinson H,Harris LW.The NHGRI-EBI GWASCatalogue:knowledgebase and deposition resource.Nucleic Acids Res.2022Nov 9:gkac1010.doi:10.1093 / nar / gkac1010.Epub ahead of print.PMID:36350656.]IEU Report[The variant call format provides efficient and robust storage of GWASummary statistics.Matthew Lyon,Shea J Andrews,Ben Elsworth,Tom R Gaunt,Gibran Hemani,Edoardo Marcora.bioRxiv doi: https: / / doi.org / 10.1101 / 2020.05.29.115824].

[0076] Introduction,R and TwoSampleMR[Hemani G,Zheng J,Elsworth B,Wade KH,HaberlandV,Baird D,Laurin C,Burgess S,Bowden J,Langdon R et al:The MR-Base platformsupports systematic causal inference across the human phenomenon.Elife 2018,7.] .

[0077] It should be understood that the Wald ratio method [Wald A: The fitting of straight lines if both variables are subject to error. The annals of mathematical statistics 1940,11(3):284-300.].

[0078] It should be understood that the IVW (inverse-variance weighted) method of Mendelian randomization [Burgess S, Butterworth A, Thompson SG: Mendelian randomization analysis with multiple genetic variants using summarized data. Genet Epidemiol 2013, 37(7): 658-665.].

[0079] It should be understood that the outcome data of psychiatric diseases of interest are input, and the instrumental variables obtained through screening are used to perform a two-sample Mendelian randomization analysis with the outcome data.

[0080] It should be understood that drug-targeted Mendelian randomization is a statistical method that uses genetic variation as an instrumental variable to estimate the effect of a specific drug target on the outcome. This method is based on the theoretical framework of the two-sample Mendelian randomization method and aims to explore the relationship between drug targets and disease outcomes to overcome the low efficiency and high cost problems commonly seen in traditional drug target research.

[0081] Genetic variation is a good choice for instrumental variables because it is naturally independent of many environmental factors, which are common sources of confounding. At the same time, since the exposure explained by genetic variation precedes the outcome, the possibility of reverse causality is effectively reduced. The MR model with genetic variation as IV relies on three core assumptions (such as Figure 2 ):

[0082] MR-IV1 (association hypothesis): Genetic variant G is associated with exposure X, that is,

[0083] MR-IV2 (independence assumption): genetic variation G is independent of confounding factor U, that is, G⊥U;

[0084] MR-IV3 (exclusive hypothesis): Genetic variation G can only affect outcome Y through exposure X and not through other pathways, that is, G⊥Y|(X,U).

[0085] Furthermore, MR allows combining multiple IVs in a single analysis and can be used when only summary level data on genetic variants and traits are publicly available.

[0086] From the causal diagram, the Mendelian randomization analysis model is:

[0087]

[0088] Among them, X is the exposure and Y is the outcome, G is the genetic variation, G is also called the instrumental variable, β is the causal effect, and ε is the random error.

[0089] Furthermore, the S103: estimating the causal effects of each target gene corresponding to multiple phenotypes, performing multiple test correction, and obtaining the corrected significant target genes specifically includes:

[0090] The results of causal effect estimation were corrected for multiple testing using the Benjamini-Hochberg method [Benjamini Y, Hochberg Y (1995) Controlling the false discovery rate: a practical and powerful approach to multiple hypothesis testing. JR Stat Soc B 57: 289–300] to control the false discovery rate. After correction, targets with P < 0.05 were considered significant proteins and targets.

[0091] Furthermore, the S103: performing sensitivity analysis on the target genes that are significant after correction to obtain the target genes for the first screening specifically includes:

[0092] S103-1: Use different MR methods to estimate the causal effect again, including maximum likelihood, weighted model, weighted median, and MR-Egger regression, to test the robustness and consistency of the results (when S103-2 to S103-5 are all passed). The most robust result is that all methods are significant and the effect direction is consistent, followed by IVW method, which is significant, and the effect direction of all other methods is consistent;

[0093] S103-2: The results of causal effect estimation were tested for instrumental variable heterogeneity using Cochran's Q test. When P>0.05, there was no heterogeneity and the results of S102-2 were used as the final results. When P<0.05, there was heterogeneity and the estimation results of causal effect were based on random effect IVW model.

[0094] S103-3: Conduct horizontal pleiotropy test on the results of causal effect estimation using MR-Egger regression;

[0095] In the test with MR-Egger intercept term of 0, P>0.05 indicates that there is no horizontal pleiotropy, and the S102-2 result is the final result; if P<0.05, the estimation of causal effect is based on the MR-Egger result;

[0096] In the global test of MR-PRESSO, P>0.05 indicates that there is no horizontal pleiotropy, and the result of S102-2 is used as the final result; if P<0.05, the estimation of causal effect is based on the result after removing outliers in MR-PRESSO;

[0097] S103-4: Identify outliers in the results of causal effect estimation: Use MR-PRESSO to identify outliers. If outliers exist, the causal effect estimation is based on the results of MR-PRESSO after removing the outliers.

[0098] S103-5: A reverse causal test was performed on the results of causal effect estimation. The MR Steiger [Hemani G, Tilling K, Davey Smith G: Orienting the causal relationship between imprecisely measured traits using GWAS summary data. PLoS Genet 2017, 13(11): e1007081.] test was used to evaluate the degree of reverse causality. P<0.05 indicated that there was no reverse causal relationship and the causal effect of the target on the disease was established. If P>0.05, the instrumental variables obtained in S101-3 were subjected to Steiger Filter. After removing the instrumental variables with causal effects in the opposite direction, S102-2 analysis and subsequent MR Steiger test were performed again. If P<0.05, the causal effect of the target on the disease was established. Otherwise, the causal effect of the target on the disease was removed.

[0099] It should be understood that Cochran's Q [Burgess S, Butterworth A, Thompson SG: Mendelian randomization analysis with multiple genetic variants using summarized data. Genet Epidemiol 2013, 37(7): 658-665.].

[0100] MR-Egger regression [Bowden J, Davey Smith G, Burgess S: Mendelian randomization with invalid instruments: effect estimation and bias detection through Eggerregression. Int J Epidemiol 2015,44(2):512-525.] or MR-PRESSO [Verbanck M, ChenCY, Neale B, Do R: Detection of widespread horizontal pleiotropy in causal relationships inferred from Mendelian randomization between complex traits and diseases. Nat Genet 2018,50(5):693-698.].

[0101] It should be understood that the beneficial effect of the above steps S103 - 1 to S103 - 5 is to ensure the consistency and robustness of the results.

[0102] Furthermore, the S103: performing co-localization analysis on the target genes screened for the first time to obtain the target genes screened for the second time specifically includes:

[0103] For protein-disease associations obtained from two-sample Mendelian randomization analysis, colocalization analysis was performed to assess whether the gene and the disease shared the same causal genetic variant;

[0104] The posterior probability under hypothesis 4 (PP.H4) was determined to be > 0.7 as evidence supporting the shared causal genetic variation hypothesis. If PP.H4 > 0.7, the causal association between the target and the disease was the highest level of evidence, and if PP.H4 < 0.7, the causal association between the target and the disease was secondary evidence.

[0105] It should be understood that the results of target-disease phenotypes supported by co-localization analysis are more reliable. For significant MR results, co-localization analysis was performed using a Bayesian framework to assess whether the gene and the disease share the same causal genetic variant. If there is a shared genetic variant, it indicates that the SNP affects the disease through its effect on the gene rather than other biological pathways.

[0106] The five hypotheses (H0 to H4) are as follows:

[0107] H0 (Null Hypothesis): No causal variant has an effect on the protein or psychiatric disorder. That is, there is no evidence that any genetic variant is associated with protein levels or psychiatric disorder outcomes.

[0108] H1 (Protein-only Hypothesis): There is only one causal variant that affects protein levels but does not affect psychiatric disease outcomes. This means that the genetic variant affects protein levels but has no direct link to disease outcomes.

[0109] H2 (Disease-only Hypothesis): Only one causal variant affects the outcome of psychiatric disease without affecting protein levels. This suggests that the genetic variant directly affects the disease outcome but is not associated with protein levels.

[0110] H3 (Independent Hypothesis): There are two different causal variants, one affecting protein levels and the other independently affecting psychiatric disease outcomes. This suggests that the protein and disease outcomes are affected by different genetic variants.

[0111] H4 (Shared Hypothesis): There is a shared causal variant that affects both protein levels and psychiatric disease outcomes. This means that genetic variants indirectly affect disease outcomes by affecting protein levels, supporting the hypothesis that proteins are potential causal factors in disease.

[0112] In colocalization analysis, by comparing the posterior probabilities (PP) of these hypotheses, it is possible to assess which hypothesis is more likely to explain the observed genetic association. Typically, PP.H4 (Posterior probability under hypothesis4) greater than a certain threshold (such as 0.7) is considered evidence in favor of the H4 hypothesis, that is, there is a shared causal variant, which supports the causal inference of affecting disease outcomes through proteins.

[0113] Furthermore, S104: for the target genes screened for the second time, pathway enrichment analysis and protein interaction network analysis were performed using R package clusterProfiler [Yu G, Wang LG, Han Y, He QY: clusterProfiler: an R package for comparing biological themes among gene clusters. OMICS 2012, 16(5): 284-287.].

[0114] It should be understood that the GO database [Gene Ontology C, Aleksander SA, Balhoff J, Carbon S, Cherry JM, Drabkin HJ, Ebert D, Feuermann M, Gaudet P, Harris NL et al: The Gene Ontology knowledgebase in 2023. Genetics 2023, 224 (1).] is a collaborative project jointly participated in and maintained by the global bioinformatics community, dedicated to creating a comprehensive and standardized gene function classification system. Its goal is to provide a unified description framework for genes and proteins so that gene functions between different species can be compared and mapped to each other. Through its detailed terminology and classification system, the GO database divides the functions of genes and proteins into three categories: biological processes, molecular functions, and cellular components, thereby achieving a systematic description of complex events occurring in organisms.

[0115] Researchers use the GO database to annotate gene functions, linking experimental data and computational predictions with specific GO terms. These annotation information is constantly enriched and updated, providing valuable resources for biological research. The GO database has a wide range of applications, from gene function prediction to disease mechanism research to drug target discovery and verification. The GO database plays an important role. By integrating data from different sources, the GO database promotes in-depth interdisciplinary research and accelerates knowledge accumulation and innovation in the biomedical field. In addition, the GO database also provides a variety of tools and resources, such as the Amigo browser, which enables users to easily query and browse the GO annotation information of genes and proteins. The GO Alliance also regularly organizes training and educational activities to help new users better understand and use the GO database. With the continuous development of the biological science field, the GO database is also constantly expanding and improving to meet new research needs and challenges. Large-scale genomic studies can increase the ability to deeply understand gene functions, and the GO database is an indispensable tool in this process. Through the GO database, researchers can link gene expression data with standardized descriptions of biological processes, molecular functions, and cellular components, so as to understand the role and regulatory mechanism of genes at a more macro level. This macroscopic perspective is of great value and significance in revealing the causes of diseases, exploring potential treatment strategies, and accelerating the development of new drugs. Therefore, the GO database is not only a platform for data storage and retrieval, but also an important engine for promoting the progress of biomedical research.

[0116] The KEGG (Kyoto Encyclopedia of Genes and Genomes) pathway database [Kanehisa M, Goto S: KEGG: kyoto encyclopedia of genes and genomes. Nucleic Acids Res 2000, 28 (1): 27-30.] is a comprehensive bioinformatics resource that provides researchers with a platform for in-depth understanding of cellular processes and disease states. The KEGG database collects and integrates a large amount of biological data, allowing scientists to explore the interactions between genes and proteins in organisms and their roles in complex biological networks. The core of the KEGG database is its pathway maps, which describe biochemical processes and molecular interactions in detail. Each pathway map is carefully drawn to show detailed views of metabolic pathways, signal transduction networks, cellular processes, etc. These views include not only the molecules involved in these processes, but also the interactions and regulatory relationships between them. Pathway enrichment analysis is an important function of the KEGG database, which allows researchers to identify pathways that are significantly enriched in specific biological states or diseases. Through this analysis, researchers can discover disease-related genes and proteins, as well as their role in the occurrence and development of diseases. For example, by comparing gene expression data in healthy and disease states, genes that are upregulated or downregulated in diseases can be identified and then associated with specific biological pathways. The KEGG database also provides a disease database that specifically collects and organizes genetic information related to human diseases. This database can help researchers understand how gene variants associated with specific diseases affect biological processes and how these variants are associated with the occurrence and development of diseases.

[0117] Furthermore, S104: using ChEMBL, DrugBank, and DGIdb databases to match target genes with drugs, if there are drugs that have undergone clinical trials or have been marketed through clinical trials, the results of old drugs being used for new purposes are obtained; if there are no drugs that have undergone clinical trials or have been marketed through clinical trials, the target genes are used as target genes for guiding new drugs.

[0118] Pathway enrichment analysis and protein interaction network analysis were used to interpret significant results of Mendelian randomization and colocalization.

[0119] Protein-protein interaction (PPI) analysis: The STRING database [Szklarczyk D, Kirsch R, Koutrouli M, Nastou K, Mehryary F, Hachilif R, Gable AL, Fang T, Doncheva NT, Pyysalo S et al: The STRING database in 2023: protein-protein association networks and functional enrichment analyses for any sequenced genome of interest. Nucleic Acids Res 2023, 51(D1): D638-D646.] is a comprehensive bioinformatics tool that focuses on revealing and analyzing protein-protein interaction (PPI) networks. It provides users with a comprehensive view of protein interactions by integrating evidence from different channels, including experimental data, information in scientific literature, records in other databases, and prediction methods based on computational biology. These rich data enable STRING to provide deep insights into the roles of proteins in biological processes, especially in signal transduction, metabolic pathways, and disease-related biological processes. The experimental data of the STRING database are derived from a variety of high-throughput technologies, such as yeast two-hybrid and mass spectrometry, providing direct experimental support for protein-protein interactions. Through automated text mining technology, STRING extracts and verifies protein interaction information from scientific literature, enhancing the reliability and coverage of the database. At the same time, it integrates protein interaction data from multiple databases such as BioGRID, IntAct and MINT, providing users with additional reference information. STRING also includes interactions predicted based on bioinformatics methods, such as gene neighborhoods, gene fusions, co-expression, etc., which helps to discover new or unverified interactions. The STRING database provides a series of user-friendly analysis and visualization tools that enable users to easily browse, analyze and interpret protein interaction networks. These tools include network analysis, functional enrichment analysis and protein function prediction, which can help users identify key protein hubs, explore protein complexes and modules, and study protein interaction patterns in specific biological processes or diseases. The STRING database has an intuitive interface and powerful visualization tools, allowing researchers to intuitively understand the complexity of protein interactions, promote the understanding of biological complexity, and provide new insights for biomedical research.

[0120] GeneMANIA [Franz M, Rodriguez H, Lopes C, Zuberi K, Montojo J, Bader GD, Morris Q: GeneMANIA update 2018. Nucleic Acids Res 2018, 46(W1): W60-W64.] is an advanced bioinformatics tool designed to predict gene function, especially for genes whose functions are not yet clear. It analyzes a user-provided gene list and uses known gene properties and a variety of association data between genes, such as gene expression data, protein interactions, genetic linkage, biological pathway information, etc. to predict the possible functions and interactions of these genes. The power of GeneMANIA lies in its comprehensive analysis capabilities. It not only considers information from a single data source, but also integrates information from multiple data sources to improve the accuracy and reliability of predictions. In this way, GeneMANIA is able to reveal complex relationships between genes, including direct physical interactions and indirect functional connections. The prediction results of the tool are usually presented in the form of a network diagram, where nodes in the network represent genes and edges represent interactions between genes. Users can customize the analysis by adjusting different parameters, such as selecting specific data types or adjusting the threshold for interactions. GeneMANIA also provides a user-friendly interface that allows researchers to easily upload gene lists, run analyses, and interpret results. The hypotheses generated by GeneMANIA can guide subsequent experimental design and help researchers determine which genes are worthy of further study. In this way, GeneMANIA not only promotes the understanding of gene function, but also provides strong support for the discovery of new biological pathways and disease-related genes. As biomedical research continues to deepen, GeneMANIA, as a predictive tool, will be more and more widely used in gene function discovery and disease mechanism research.

[0121] Furthermore, the S105: conducting a safety evaluation on the target gene used to guide the new drug, and evaluating the safety of the target gene, specifically includes:

[0122] For the discovered targets, SAIGE was used to analyze 1403 binary disease phenotype outcomes in UKB to assess the safety of target genes;

[0123] There are summary statistics for a total of 1,403 binary phenotypes in the UKB SAIGE GWAS. To improve statistical power, phenotypes with 500 or more cases were selected, so 784 binary disease traits were used as outcomes and full phenotype group Mendelian randomization was performed.

[0124] Due to the large number of disease phenotypes analyzed, the Manhattan plot was used to display the association analysis results, with chromosomes as the horizontal axis and the log-values ​​of the P values ​​obtained from each association analysis as the horizontal axis. 10 (P) is a scatter plot with the vertical axis.

[0125] It should be understood that the Mendelian randomization analysis platform is an advanced integrated system (such as Figure 3 and Figure 4 The platform integrates high-throughput screening technology and genetic data analysis tools, allowing users to easily upload data, select analysis parameters, and automate the entire analysis process through a workflow engine.

[0126] In terms of data integration, the platform is able to process and integrate large-scale data sets from multiple bioinformatics databases, including genomic, proteomic, and drug target data. The drug target screening module uses these data to identify potential drug action sites. The Mendelian randomization analysis module uses genetic variation as an instrumental variable to evaluate the possible causal relationship between drug targets and diseases. The colocalization analysis module further verifies whether these genetic variations affect the disease through specific biological pathways.

[0127] The platform also includes pathway enrichment and PPI analysis modules, which combine databases such as GO, KEGG, STRING and GeneMANIA to deeply analyze protein functions and interaction networks. In addition, the drug targeting assessment module uses databases such as DGIdb, DrugBank and ChEMBL to provide in-depth insights into drug development prospects (such as Figure 5 and Figure 6 shown).

[0128] To enhance the interpretability of the results, the platform's automated report generation system can output detailed reports containing analysis results, charts, and conclusions. The data management and quality control modules ensure the accuracy and reliability of data during the analysis process, while the platform implements data encryption and access control to protect data security and user privacy.

[0129] In terms of technical architecture, the platform adopts an extensible and compatible design to support the integration of new tools and new data types in the future. The platform provides remote access and server computing resource allocation functions to ensure global availability. The platform will continue to optimize the code to improve analysis speed and computing performance.

[0130] Based on the drug-targeted Mendelian randomization method, the multi-omics database of mental illness and the drug-targeted Mendelian randomization platform were developed to realize an integrated analysis process from determining the mental illness of interest to selecting the database of interest for drug-targeted Mendelian randomization analysis and sensitivity analysis. Mental illnesses include major depression, bipolar disorder, neuroticism, post-traumatic stress disorder, schizophrenia, anorexia nervosa, attention deficit hyperactivity disorder, autism spectrum disorder, insomnia and other diseases. The database includes eQTL data from eqtlgen and GTEx, pQTL data from ARIC study and UKB-PPP, and GWAS data from GWAS catalog, ieu and Finngen. According to the function of the drug, the drug is divided into drugs for lowering low-density lipoprotein cholesterol, lowering triglycerides, lowering systolic blood pressure and other functional drugs. The database contains the target protein of the drug, the corresponding gene encoding the protein and its location on the chromosome. Through the operational platform interface built by R shiny, the commonly used methods of Mendelian randomization are integrated, and the results are output in the form of data files, visual graphics, and interpretable reports, realizing the automation of drug-targeted Mendelian randomization analysis.

[0131] Through the establishment of an operational platform, a technological breakthrough has been achieved, which has improved the operability of drug-targeted Mendelian randomization analysis. At the same time, the establishment of an operational platform has also provided clinicians and drug development companies with many new ideas for drug development. By using eQTL, pQTL and GWAS data, researchers can explore the reuse of marketed drugs, discover new drugs for treating diseases, and explore the targets and mechanisms of action of drugs. In addition, the platform can also provide researchers with the latest, large-sample eQTL, pQTL and GWAS data and their genetic information to help them gain a deeper understanding of the drug mechanism of action and the pathogenesis of the disease, and explore possible biological mechanisms to obtain causal pathways. The platform simplifies the experimental process of drug-targeted Mendelian randomization analysis, while improving the accuracy and efficiency of the analysis, which not only helps to accelerate the drug development process, but also provides new possibilities for the innovation of treatment strategies.

[0132] The drug-targeted Mendelian randomization analysis platform of the present invention has the following advantages: Clarify the causal relationship between drugs and diseases: Through the Mendelian randomization method, the causal relationship between drugs and diseases can be effectively distinguished, providing a reliable basis for drug target selection. Integrate multi-omics data and simplify the analysis process: The platform integrates a variety of omics data and analysis tools, simplifies the data integration and analysis process, and improves research efficiency. Promote drug reuse and new drug development: Through data-driven analysis methods, new indications for marketed drugs are discovered to promote drug reuse and new drug development.

[0133] Embodiment 2

[0134] This embodiment provides a system for targeted analysis of psychiatric drugs based on Mendelian randomization, comprising:

[0135] An acquisition module is configured to: acquire an existing proteomic dataset and an existing transcriptomic dataset; perform preprocessing operations on the existing proteomic dataset and the existing transcriptomic dataset to obtain a preprocessed proteomic dataset and a preprocessed transcriptomic dataset; acquire a target mental illness dataset, an existing drug database, and an existing phenotype database;

[0136] An analysis module is configured to: perform analysis and processing using a Mendelian randomization analysis process based on a target mental illness dataset, an existing drug database, an existing phenotype database, a preprocessed proteome dataset, and a preprocessed transcriptome dataset to obtain causal effect estimates of each target gene corresponding to multiple phenotypes;

[0137] The screening module is configured as follows: for each target gene, causal effect estimation corresponding to multiple phenotypes is performed, multiple test correction is performed, and the target genes that are significant after correction are obtained; for the target genes that are significant after correction, sensitivity analysis is performed to obtain the target genes for the first screening; for the target genes for the first screening, co-localization analysis is performed to obtain the target genes for the second screening;

[0138] The output module is configured as follows: for the target genes screened for the second time, pathway enrichment analysis is first performed, followed by protein interaction network analysis, and finally drug matching is performed on the target genes of the protein interaction network to obtain the prediction results of old drug new use or use the target genes as target genes for guiding the preparation of new drugs. Old drug new use means using drugs already available on the market to treat target mental illnesses.

[0139] It should be noted that the acquisition module, analysis module, screening module and output module described above correspond to steps S101 to S104 in Embodiment 1, and the examples and application scenarios implemented by the modules and the corresponding steps are the same, but are not limited to the contents disclosed in Embodiment 1. It should be noted that the modules described above as part of the system can be executed in a computer system such as a set of computer executable instructions.

[0140] The description of each embodiment in the above embodiments has different emphases. For parts not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0141] The proposed system can be implemented in other ways. For example, the system embodiment described above is only illustrative, and the division of the modules is only a logical function division. In actual implementation, there may be other division methods, such as multiple modules can be combined or integrated into another system, or some features can be ignored or not executed.

[0142] Embodiment 3

[0143] This embodiment also provides an electronic device, including: one or more processors, one or more memories, and one or more computer programs; wherein the processor is connected to the memory, and the one or more computer programs are stored in the memory. When the electronic device is running, the processor executes the one or more computer programs stored in the memory so that the electronic device executes the method described in the above embodiment one.

[0144] It should be understood that in this embodiment, the processor may be a central processing unit CPU, and the processor may also be other general-purpose processors, digital signal processors DSP, application-specific integrated circuits ASIC, off-the-shelf programmable gate arrays FPGA or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.

[0145] The memory may include a read-only memory and a random access memory, and provide instructions and data to the processor. A portion of the memory may also include a non-volatile random access memory. For example, the memory may also store information about the device type.

[0146] In the implementation process, each step of the above method can be completed by an integrated logic circuit of hardware in a processor or an instruction in the form of software.

[0147] The method in the first embodiment can be directly embodied as a hardware processor, or a combination of hardware and software modules in the processor. The software module can be located in a mature storage medium in the field such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory, or an electrically erasable programmable memory, a register, etc. The storage medium is located in the memory, and the processor reads the information in the memory and completes the steps of the above method in combination with its hardware. To avoid repetition, it will not be described in detail here.

[0148] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with this embodiment can be implemented in electronic hardware or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0149] Embodiment 4

[0150] This embodiment further provides a computer-readable storage medium for storing computer instructions. When the computer instructions are executed by a processor, the method described in the first embodiment is completed.

[0151] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A targeted analysis method for drugs for mental illness based on Mendelian randomization, characterized by: include: Obtain existing proteome datasets and transcriptome datasets; Preprocessing the existing proteomic dataset and the existing transcriptomic dataset to obtain a preprocessed proteomic dataset and a preprocessed transcriptomic dataset; Obtain target mental illness datasets, existing drug databases, and existing phenotype databases; Based on the target mental illness dataset, existing drug database, existing phenotype database, pre-processed proteome dataset and pre-processed transcriptome dataset, the Mendelian randomization analysis process is used to analyze and process, and the causal effect estimation of each target gene corresponding to multiple phenotypes is obtained; For each target gene, the causal effect of multiple phenotypes is estimated, and multiple test correction is performed to obtain the target genes that are significant after correction; for the target genes that are significant after correction, sensitivity analysis is performed to obtain the target genes for the first screening; for the target genes for the first screening, co-localization analysis is performed to obtain the target genes for the second screening; For the target genes screened for the second time, pathway enrichment analysis is first performed, then protein interaction network analysis is performed, and finally drug matching is performed on the target genes in the protein interaction network to obtain prediction results for new uses of old drugs or use the target genes as target genes to guide the preparation of new drugs.

2. The method for targeted analysis of psychiatric drugs based on Mendelian randomization as claimed in claim 1, characterized in that: The method also includes: conducting a safety evaluation on the target gene used to guide the preparation of new drugs to assess the safety of the target gene: for the discovered targets, using SAIGE to analyze 1403 binary disease phenotype outcomes in UKB to assess the safety of the target gene.

3. The method for targeted analysis of psychiatric drugs based on Mendelian randomization according to claim 1, characterized in that: The preprocessing operation is performed on the existing proteomic dataset and the existing transcriptomic dataset to obtain the preprocessed proteomic dataset and the preprocessed transcriptomic dataset. The preprocessing operation specifically includes: (1-1) Determine the significance of the whole genome P < 5e -8 Is it true? If so, the single nucleotide polymorphism is considered as an instrumental variable for subsequent processing; if not, the single nucleotide polymorphism is not used as an instrumental variable; (1-2) screening cis-pQTL from the single nucleotide polymorphisms obtained in (1-1); cis-pQTL is defined as a protein-coding gene within 1 Mb from 500 kb upstream of the start site to 500 kb downstream of the stop site, or a protein-coding gene within 1 Mb from 500 kb upstream and downstream of the transcription start site; (1-3) Cluster the single nucleotide polymorphisms obtained by (1-2) by linkage disequilibrium, and the specific clustering parameter is r 2 = 0.01 with a window of 1000 kb and the 1000Genomes Project Phase 3 as the reference panel to ensure the independence of the selected SNPs: r 2 = 0.01, the window is 1Mb, remove the r 2 >0.01 SNPs.

4. The method for targeted analysis of psychiatric drugs based on Mendelian randomization as claimed in claim 3, characterized in that: Based on the target mental illness dataset, existing drug database, existing phenotype database, pre-processed proteome dataset and pre-processed transcriptome dataset, the Mendelian randomization analysis process was used to analyze and process, and the causal effect estimates of each target gene corresponding to multiple phenotypes were obtained, including: (2-1) From the GWAS-CATALOG database, GWAS data of mental illnesses were selected as outcome data, including major depressive disorder, attention deficit hyperactivity disorder, post-traumatic stress disorder, schizophrenia, autism spectrum disorder, anorexia nervosa, bipolar disorder, neuroticism, and anxiety; (2-2) Two-sample Mendelian randomization analysis was performed to estimate the causal effect of each target gene on multiple psychiatric disorders. The input files of the two-sample Mendelian randomization analysis included: the instrumental variables of the target gene and the outcomes of the two-sample Mendelian randomization analysis. The instrumental variables of the target gene were the SNPs selected from (1-1) to (1-3), and the outcomes of the two-sample Mendelian randomization analysis were the GWAS data of psychiatric disorders selected from (2-1). When there was only one instrumental variable, the Wald ratio method of Mendelian randomization was used to estimate the causal effect. When there were two or more instrumental variables, the IVW method of Mendelian randomization was used to estimate the causal effect.

5. The method for targeted analysis of psychiatric drugs based on Mendelian randomization according to claim 1, characterized in that: For each target gene, the causal effect of multiple phenotypes is estimated, and multiple testing correction is performed to obtain the target genes that are significant after correction, including: The results of causal effect estimation were corrected for multiple tests using the Benjamini-Hochberg method to control the false discovery rate. After correction, targets with P < 0.05 were considered significant proteins and targets.

6. The method for targeted analysis of psychiatric drugs based on Mendelian randomization as claimed in claim 4, characterized in that: Sensitivity analysis was performed on the target genes that were significant after correction to obtain the target genes for the first screening, including: (3-1): The causal effect was estimated again using different MR methods, including maximum likelihood, weighted model, weighted median, and MR-Egger regression, to test the robustness and consistency of the results. The most robust result was that all methods were significant and the effect direction was consistent, followed by the IVW method, which was significant, and the effect direction of all other methods was consistent; (3-2): The results of causal effect estimation were tested for instrumental variable heterogeneity using Cochran's Q test. When P>0.05, there was no heterogeneity and the result of (2-2) was used as the final result. When P<0.05, there was heterogeneity and the estimation result of causal effect was based on random effect IVW model. (3-3): The results of causal effect estimation were tested for horizontal pleiotropy using MR-Egger regression. In the test with MR-Egger intercept term of 0, if P>0.05, it means that there is no horizontal pleiotropy, and the result of (2-2) is the final result. If P<0.05, the causal effect estimation is based on the result of MR-Egger. In the global test of MR-PRESSO, if P>0.05, it means that there is no horizontal pleiotropy, and the result of (2-2) is the final result. If P<0.05, the causal effect estimation is based on the result after removing outliers in MR-PRESSO. (3-4): Identify outliers in the results of causal effect estimation: Use MR-PRESSO to identify outliers. If outliers exist, the causal effect estimation is based on the results of MR-PRESSO after removing the outliers. (3-5): A reverse causal test is performed on the results of causal effect estimation, and the MR Steiger test is used to evaluate the degree of reverse causality. If P < 0.05, it means that there is no reverse causal relationship, and the causal effect of the target on the disease is established. If P > 0.05, the instrumental variables obtained in (1-3) are subjected to Steiger Filter. After eliminating the instrumental variables with causal effects in the opposite direction, (2-2) analysis and subsequent MR Steiger test are performed again. If P < 0.05, the causal effect of the target on the disease is established. Otherwise, the causal effect of the target on the disease is eliminated.

7. The method for targeted analysis of psychiatric drugs based on Mendelian randomization as claimed in claim 1, characterized in that: The target genes screened in the first screening were subjected to co-localization analysis to obtain the target genes for the second screening, including: For the protein-disease association obtained by two-sample Mendelian randomization analysis, co-localization analysis was performed to evaluate whether the gene and the disease shared the same causal genetic variation; the posterior probability of hypothesis H4 PP.H4>0.7 was determined as evidence supporting the hypothesis of shared causal genetic variation; if PP.H4>0.7, the causal association between the target and the disease was the highest level of evidence, and if PP.H4<0.7, the causal association between the target and the disease was secondary evidence.

8. A Mendelian randomization-based drug targeting analysis system for mental illness, characterized by: include: An acquisition module is configured to: acquire an existing proteome dataset and an existing transcriptome dataset; Preprocessing the existing proteomic dataset and the existing transcriptomic dataset to obtain a preprocessed proteomic dataset and a preprocessed transcriptomic dataset; Obtain target mental illness datasets, existing drug databases, and existing phenotype databases; An analysis module is configured to: perform analysis and processing using a Mendelian randomization analysis process based on a target mental illness dataset, an existing drug database, an existing phenotype database, a preprocessed proteome dataset, and a preprocessed transcriptome dataset to obtain causal effect estimates of each target gene corresponding to multiple phenotypes; The screening module is configured as follows: for each target gene, causal effects corresponding to multiple phenotypes are estimated, multiple test correction is performed, and significant target genes are obtained after correction; for the target genes that are significant after correction, sensitivity analysis is performed to obtain target genes for the first screening; Conduct co-localization analysis on the target genes screened for the first time to obtain the target genes for the second screening; The output module is configured as follows: for the target genes screened for the second time, pathway enrichment analysis is first performed, followed by protein interaction network analysis, and finally drug matching is performed on the target genes of the protein interaction network to obtain prediction results for new uses of old drugs or to use the target genes as target genes for guiding the preparation of new drugs.

9. An electronic device, comprising: a memory for non-transitory storage of computer readable instructions; as well as a processor for executing the computer readable instructions, When the computer-readable instructions are executed by the processor, the method described in any one of claims 1 to 7 is executed.

10. A storage medium, characterized in that: The computer-readable instructions are non-transitory stored, wherein when the non-transitory computer-readable instructions are executed by a computer, the instructions of the method according to any one of claims 1 to 7 are executed.

Images