Predictive device for antipsychotic drug efficacy

Abstract

The present invention discloses a device for predicting the efficacy of antipsychotic drugs. The method processes and calculates the volume of a subcortical structure of a subject through a variation determination module, calculates the correlation coefficient between the volumes of two brain regions in the volume, and determines the degree of variation of the covariant connection of the subcortical structure according to the correlation coefficient; a gene acquisition module obtains the whole genome gene expression value of the autopsy brain tissue corresponding to the subcortical structure, and determines the gene of interest according to the whole genome gene expression value and the degree of variation; a prediction module obtains the methylation level corresponding to the gene of interest, constructs a stepwise multivariate linear regression model according to the methylation level, and obtains the prediction result of the efficacy of antipsychotic drugs through the multivariate linear regression model, which can effectively predict the efficacy of antipsychotic drugs, improve the efficiency of the prediction of the efficacy of antipsychotic drugs, and ensure the accuracy of the prediction of the efficacy of antipsychotic drugs.

Description

Technical Field

[0001] The invention relates to the field of biotechnology, and in particular to a device for predicting the efficacy of antipsychotic drugs. Background Art

[0002] Approximately 70% of first-episode schizophrenia patients subsequently develop treatment resistance, having shown a lack of response to non-clozapine antipsychotic treatment from their first presentation to mental health services; early treatment response is considered one of the reliable indicators for predicting patients' clinical outcomes, but it is variable and cannot be accurately predicted by physicians; therefore, it is critical to identify biomarkers with predictive potential by studying their relationship with early treatment response in schizophrenia.

[0003] Currently, there is still a lack of effective methods for constructing subcortical structural covariant connectivity and DNA methylation multimodal data for application in the prediction of antipsychotic drug efficacy. Summary of the invention

[0004] The main purpose of the present invention is to provide a device for predicting the efficacy of antipsychotic drugs, aiming to solve the technical problems in the prior art that antipsychotic drug treatment lacks response and the data effect cannot be accurately predicted by doctors.

[0005] In a first aspect, the present invention provides a device for predicting the efficacy of an antipsychotic drug, the device comprising:

[0006] a variation determination module, for processing and calculating the volume of the subject's subcortical structure, calculating the correlation coefficient between the volumes of two brain regions in the volume, and determining the degree of variation of the covariant connection of the subcortical structure according to the correlation coefficient;

[0007] A gene acquisition module, used to acquire the whole genome gene expression value of the autopsy brain tissue corresponding to the subcortical structure, and determine the gene of interest according to the whole genome gene expression value and the degree of variation;

[0008] The prediction module is used to obtain the methylation level corresponding to the gene of interest, construct a stepwise multiple linear regression model according to the methylation level, and obtain the prediction result of the efficacy of antipsychotic drugs through the multiple linear regression model.

[0009] Optionally, the variation determination module is further used to process and calculate brain MRI scan data of the subject, and determine the volume of the subcortical structure based on the brain MRI scan data;

[0010] The variation determination module is further used to calculate the correlation coefficient between the volumes of the two brain regions in the volume using the Pearson correlation analysis method, taking each pair of brain regions as an edge and using the correlation coefficient as the covariate connection strength of each edge;

[0011] The variation determination module is further used to obtain the covariant connection strength difference of the covariant connection strength of each edge, and determine the variation degree of the covariant connection of the subcortical structure according to the covariant connection strength difference.

[0012] Optionally, the variation determination module is also used to process and calculate the subject's brain magnetic resonance imaging data, perform head motion correction, skull stripping, coordinate transformation, and subcortical white matter and dark gray matter volume structure segmentation on the magnetic resonance imaging data, and obtain the volume of the subcortical structure.

[0013] Optionally, the variation determination module is further used to obtain symptom relief data of the subject after treatment, and divide the covariate connection strength into a response group and a non-response group according to the symptom relief data;

[0014] The variation determination module is also used to use a Z test to compare the difference in covariant connection strength of each edge in the response group and the non-response group, obtain the Z value of each edge, form a Z statistical graph based on the Z value, and determine the degree of variation of the covariant connection of the subcortical structure through the Z statistical graph.

[0015] Optionally, the gene acquisition module is further used to obtain the whole genome gene expression value of the autopsy brain tissue corresponding to the subcortical structure in a preset database;

[0016] The gene acquisition module is further used to calculate the correlation value between the whole genome gene expression value and the variation degree by using the least square method, and obtain the gene whose correlation value is greater than a preset correlation threshold as the target gene;

[0017] The gene acquisition module is further used to take the target gene and the intersection gene in the preset risk gene set, and take the intersection gene as the gene of interest.

[0018] Optionally, the gene acquisition module is further used to acquire original gene expression data of autopsy brain tissue corresponding to the subcortical structure in a preset database;

[0019] The gene acquisition module is also used to verify the gene annotation of the original gene expression data through probes, filter out probes that do not exceed background noise, and select representative target probes;

[0020] The original gene expression data is atlas-aligned by the target probe, and the expression value is normalized to obtain the whole genome gene expression value of the autopsy brain tissue.

[0021] Optionally, the prediction module is further used to analyze the methylation data of the whole genome DNA of the peripheral blood of the subject, and obtain the methylation level of the CpG dinucleotide site corresponding to the gene of interest from the methylation data;

[0022] The prediction module is further used to obtain symptom change data of the subject after treatment, and to construct a stepwise multivariate linear regression model according to the methylation level and the symptom change data;

[0023] The prediction module is also used to obtain the prediction results of the efficacy of antipsychotic drugs through the multivariate linear regression model.

[0024] Optionally, the prediction module is also used to analyze the methylation data of the whole genome DNA of the subject's peripheral blood, obtain polymorphic sites in the CpG dinucleotide sites from the methylation data, eliminate the polymorphic sites, obtain the methylation risk index values ​​of the remaining CpG sites, and obtain the average value of the methylation values ​​corresponding to the methylation risk index value from the gene of interest as the methylation level.

[0025] Optionally, the prediction module is also used to obtain symptom change data of the subject after treatment, use the symptom change data as the dependent variable of a preset linear regression equation, use the methylation level as the independent variable in the preset linear regression equation, and construct a stepwise multivariate linear regression model based on the dependent variable and the independent variable.

[0026] Optionally, the prediction module is also used to evaluate the reduction rate of the symptom change data to obtain an evaluation result, perform a residual normality test on the multivariate linear regression model based on the evaluation result, and use the efficacy data that passes the test as the prediction result of the efficacy of the antipsychotic drug.

[0027] The device for predicting the efficacy of antipsychotic drugs proposed in the present invention processes and calculates the volume of the subcortical structure of the subject through a variation determination module, calculates the correlation coefficient between the volumes of two brain regions in the volume, and determines the degree of variation of the covariant connection of the subcortical structure according to the correlation coefficient; the gene acquisition module obtains the whole genome gene expression value of the autopsy brain tissue corresponding to the subcortical structure, and determines the gene of interest according to the whole genome gene expression value and the degree of variation; the prediction module obtains the methylation level corresponding to the gene of interest, constructs a stepwise multivariate linear regression model according to the methylation level, and obtains the prediction result of the efficacy of antipsychotic drugs through the multivariate linear regression model, which can effectively predict the efficacy of antipsychotic drugs, improve the efficiency of the prediction of the efficacy of antipsychotic drugs, and ensure the accuracy of the prediction of the efficacy of antipsychotic drugs. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] Figure 1 This is a functional module diagram of the first embodiment of the device for predicting the efficacy of antipsychotic drugs according to the present invention.

[0029] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0030] It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention.

[0031] The solution of the embodiment of the present invention is mainly: processing and calculating the volume of the subcortical structure of the subject through a variation determination module, calculating the correlation coefficient of the volume of the two brain regions in the volume, and determining the variation degree of the covariant connection of the subcortical structure according to the correlation coefficient; the gene acquisition module obtains the whole genome gene expression value of the autopsy brain tissue corresponding to the subcortical structure, and determines the gene of interest according to the whole genome gene expression value and the variation degree; the prediction module obtains the methylation level corresponding to the gene of interest, constructs a stepwise multivariate linear regression model according to the methylation level, and obtains the prediction result of the efficacy of antipsychotic drugs through the multivariate linear regression model, which can effectively predict the efficacy of antipsychotic drugs, improve the efficiency of the prediction of the efficacy of antipsychotic drugs, ensure the accuracy of the prediction of the efficacy of antipsychotic drugs, and solve the technical problem in the prior art that antipsychotic drugs lack response and the data effect cannot be accurately predicted by doctors.

[0032] The invention provides a device for predicting the efficacy of antipsychotic drugs.

[0033] Reference Figure 1 , Figure 1 This is a functional module diagram of the first embodiment of the device for predicting the efficacy of antipsychotic drugs according to the present invention.

[0034] In a first embodiment of the device for predicting the efficacy of antipsychotic drugs of the present invention, the device for predicting the efficacy of antipsychotic drugs comprises:

[0035] The variation determination module 10 is used to process and calculate the volume of the subject's subcortical structure, calculate the correlation coefficient between the volumes of two brain regions in the volume, and determine the degree of variation of the covariant connection of the subcortical structure based on the correlation coefficient.

[0036] It should be noted that the volume of the subcortical structure of the subject's brain is obtained, and then the correlation coefficient between the volumes of any two brain regions in the volume can be calculated, and then the degree of variation of the covariant connection of the subcortical structure can be determined through the correlation coefficient.

[0037] It is understandable that the 14 subcortical structures of the brain generally include: bilateral thalamus, bilateral amygdala, bilateral hippocampus, bilateral caudate nucleus, bilateral putamen, bilateral globus pallidus and bilateral nucleus accumbens. Each of these 14 brain regions is defined by the automatic segmentation program of the FreeSurfer toolkit and extracted from the bilateral brain. The brain size of each subject is different, so the present invention calculates the total intracranial volume of each subject, and the total intracranial volume of each subject is included as a control variable when performing inter-group difference analysis on the volume of the subcortical structure.

[0038] Furthermore, the variation determination module 10 is also used to process and calculate the brain MRI scan data of the subject, and determine the volume of the subcortical structure based on the brain MRI scan data.

[0039] The variation determination module 10 is further used to calculate the correlation coefficient between the volumes of the two brain regions in the volume using the Pearson correlation analysis method, taking each pair of brain regions as an edge and using the correlation coefficient as the covariate connection strength of each edge.

[0040] The variation determination module 10 is further used to obtain the covariant connection strength difference of the covariant connection strength of each edge, and determine the variation degree of the covariant connection of the subcortical structure according to the covariant connection strength difference.

[0041] It should be noted that the brain data of subjects (schizophrenia patients) scanned on an MRI scanner can generally be preprocessed using the FreeSurfer toolkit to obtain the volume of the subcortical structure (including 14 brain regions).

[0042] It can be understood that the Pearson correlation analysis method can be used to calculate the correlation coefficient between the volumes of pairwise brain regions in the subcortical structure volume. Each pair of brain regions is called an "edge", and the correlation coefficient is used as the covariate connection strength of each "edge". In actual operation, the subjects can be divided into a treatment responder group and a non-responder group according to the symptom relief after 8 weeks of treatment, and the Z test can be used to compare the inter-group differences in the covariate connection strength of the subcortical structure between the responder group and the non-responder group.

[0043] In the specific implementation, the Pearson correlation analysis method was used to calculate the correlation coefficient between the volumes of two brain regions in the subcortical structure volume. Each pair of brain regions was called an "edge", and the correlation coefficient was used as the covariant connection strength of each "edge". According to the symptom relief of the subjects after 8 weeks of treatment, they were divided into a treatment response group and a non-response group. The Z test was used to compare the between-group differences in the covariant connection strength of the subcortical structure between the response group and the non-response group. It was found that: compared with the non-responders, the right thalamus and the right globus pallidus were significantly higher in the responders (P = 0.00 3, Z = 3.99, FDR corrected) and between the left thalamus and the left globus pallidus (P = 0.002, Z = 4.20, FDR corrected) were significantly increased, while the covariant connectivity between the left thalamus and the right hippocampus (P = 0.021, Z = -3.39, FDR corrected) and between the right putamen and the right caudate nucleus were significantly lower (P = 0.023, Z = -3.29, FDR corrected); where P is the significance level in the Z test, Z is the standard score in the test, and FDR is the False Discovery Rate (FDR). If the P value is less than 0.05, it means that the difference between the two samples is significant.

[0044] It should be understood that the strength difference of the covariant connection strength of each edge can be obtained, so that the degree of variation of the covariant connection of the subcortical structure can be determined.

[0045] Furthermore, the variation determination module 10 is also used to process and calculate the brain MRI scan data of the subject, perform head motion correction, skull stripping, coordinate transformation and subcortical white matter and dark gray matter volume structure segmentation on the MRI scan data, and obtain the volume of the subcortical structure.

[0046] In a specific implementation, the brain of a subject (i.e., a schizophrenia patient) is scanned on an MRI scanner to obtain the MRI scan data, and the subcortical structure volume is calculated on the MRI scan data, that is, the MRI scan data is image processed, and the image processing includes head motion correction, skull stripping, automatic coordinate transformation, and segmentation of subcortical white matter and dark gray matter volume structures; in the automatic segmentation process, FreeSurfer automatically marks the subcortical structure category based on the atlas-based Bayesian segmentation process, and the FreeSurfer analysis process is performed without any human intervention.

[0047] Furthermore, the variation determination module 10 is also used to obtain symptom relief data of the subject after treatment, and divide the covariate connection strength into a response group and a non-response group according to the symptom relief data;

[0048] The variation determination module 10 is also used to use a Z test to compare the difference in covariant connection strength of each edge in the response group and the non-response group, obtain the Z value of each edge, form a Z statistical graph based on the Z value, and determine the degree of variation of the covariant connection of the subcortical structure through the Z statistical graph.

[0049] It is understandable that the Z test is generally used to compare the difference in the covariant connection strength of each "edge" between the response group and the non-response group, and the Z values ​​of all "edges" form a Z statistical map. The Z statistical map is used to represent the degree of variation in the covariant connection of subcortical structures related to treatment response.

[0050] In the specific implementation, a "7×7" matrix can be constructed, that is, 7 left subcortical regions × 7 left subcortical regions, and 21 non-repeating "edges" are selected from the "7×7" matrix; then the Z test is used to compare the differences in covariant connections between the treatment response group and the non-response group in the 21 "edges". In this process, the corresponding z value can be obtained for each "edge", and finally a Z statistical graph composed of 21 Z values ​​is obtained; the Z statistical graph is used to represent the degree of variation in the covariant connections of the subcortical structures related to the treatment response.

[0051] The gene acquisition module 20 is used to obtain the whole genome gene expression value of the autopsy brain tissue corresponding to the subcortical structure, and determine the gene of interest according to the whole genome gene expression value and the degree of variation.

[0052] It is understandable that by obtaining the whole genome gene expression value of the autopsy brain tissue corresponding to the subcortical structure, and correlating the whole genome gene expression value with the degree of variation, the corresponding gene of interest can be obtained.

[0053] Furthermore, the gene acquisition module 20 is also used to obtain the whole genome gene expression value of the autopsy brain tissue corresponding to the subcortical structure in a preset database.

[0054] The gene acquisition module 20 is further used to calculate the correlation value between the whole genome gene expression value and the variation degree by using the least square method, and obtain the gene whose correlation value is greater than a preset correlation threshold as the target gene.

[0055] The gene acquisition module 20 is further used to take the target gene and the intersection gene in the preset risk gene set as the gene of interest.

[0056] In a specific implementation, the preset database is a pre-set database, which may be the Allen brain bank, and the whole genome gene expression values ​​of the autopsy brain tissue corresponding to the subcortical structure may be obtained in the Allen brain bank; and then the least squares method is used to calculate the correlation between the whole genome gene expression values ​​of the autopsy brain tissue corresponding to the subcortical structure and the obtained Z statistical graph, to obtain genes that are significantly correlated with the Z statistical graph, and then the gene is intersected with a preset risk gene set, such as the 108 risk loci for schizophrenia that have been publicly published. Of course, it may also be other risk gene sets, and this embodiment is not limited to this, and the intersection gene is the gene of interest.

[0057] It can be understood that the present embodiment uses partial least squares (PLS) regression to detect the spatial correlation between the gene expression value of the "edge" and the Z statistic map; gene expression data is used as the predictor variable and the Z statistic map is used as the response variable; the first component (PLS1) in PLS is a linear combination of the gene expression values ​​with the strongest association with the Z statistic map; a permutation test (1000 times) is used to test the null hypothesis that the association between the brain expression value and the Z statistic map explained by PLS1 is no more than expected by chance; the Z value is calculated by the ratio of the weight of the gene expression in each region to its Bootstrap standard error; after FDR correction (P<0.05), the present invention obtains a gene set that reliably contributes to PLS1, which contains a total of 991 genes, and the TOP20 genes are shown in Table 1 as an illustration for subsequent analysis, and Table 1 is as follows:

[0058] Table 1 Schematic representation of the TOP20 genes in PLS1

[0059]

[0060]

[0061] Furthermore, the gene acquisition module 20 is also used to acquire original gene expression data of autopsy brain tissue corresponding to the subcortical structure in a preset database.

[0062] The gene acquisition module is also used to verify the gene annotation of the original gene expression data through probes, filter out probes that do not exceed background noise, and select representative target probes.

[0063] The original gene expression data is atlas-aligned by the target probe, and the expression value is normalized to obtain the whole genome gene expression value of the autopsy brain tissue.

[0064] In the specific implementation, transcription profiles were obtained from the Allen brain database, including expression data of 20,737 genes represented by 58,692 probes; the original gene expression data were preprocessed, and the main steps were as follows: (i) verifying the annotation of probes to genes; (ii) filtering probes that did not exceed background noise; (iii) selecting representative probes; (iv) aligning the samples to the DK 68 atlas; (v) normalizing the expression values; after preprocessing, each brain tissue had 10,027 probes corresponding to 10,027 genes.

[0065] It is understandable that in order to extract autopsy brain tissue samples spatially located in the subcortical structure, this embodiment only selects (1) samples with structural annotations corresponding to the subcortical structure provided by AHBA, and (2) samples whose MNI coordinates can be accurately mapped to the subcortical structure; the average expression level of each gene in all brain tissue samples in a brain region is taken as the expression value of each gene in the brain region; the average expression level of each gene in the two brain regions connected by each covariantly connected "edge" is defined as the expression level of the gene in the "edge" for subsequent analysis.

[0066] The prediction module 30 is used to obtain the methylation level corresponding to the gene of interest, construct a stepwise multiple linear regression model according to the methylation level, and obtain the prediction result of the efficacy of the antipsychotic drug through the multiple linear regression model.

[0067] It should be understood that after obtaining the methylation data corresponding to the gene of interest, the corresponding methylation level can be obtained, and a stepwise multivariate linear regression model can be constructed through the methylation level. The stepwise multivariate linear regression model is a data model in which different methylation levels correspond to different antipsychotic drug therapeutic effects. The prediction results of predicting the efficacy of antipsychotic drugs are obtained through the multivariate linear regression model.

[0068] Furthermore, the prediction module 30 is also used to analyze the methylation data of the whole genome DNA of the peripheral blood of the subject, and obtain the methylation level of the CpG dinucleotide site corresponding to the gene of interest from the methylation data.

[0069] The prediction module 30 is further used to obtain the symptom change data of the subject after treatment, and to construct a stepwise multivariate linear regression model according to the methylation level and the symptom change data.

[0070] The prediction module 30 is also used to obtain the prediction result of the efficacy of antipsychotic drugs through the multivariate linear regression model.

[0071] It should be noted that the whole genome DNA methylation data of the peripheral blood of the subjects is obtained, and the methylation levels of the CpG dinucleotide sites corresponding to the genes of interest are extracted from the whole genome methylation samples of the peripheral blood of the patients. Then, the stepwise multivariate linear regression analysis method can be used to analyze the association between the DNA methylation of the genes of interest of the subjects and the symptom change data, such as the symptom changes after 8 weeks of treatment.

[0072] Furthermore, the prediction module 30 is also used to analyze the methylation data of the whole genome DNA of the subject's peripheral blood, obtain the polymorphic sites in the CpG dinucleotide sites from the methylation data, eliminate the polymorphic sites, obtain the methylation risk index values ​​of the remaining CpG sites, and obtain the average value of the methylation values ​​corresponding to the methylation risk index value from the gene of interest as the methylation level.

[0073] In a specific implementation, the Illumina 450K methylation chip can be used to detect the methylation level of CpG sites in the whole genome of the subject's peripheral blood, and the sites with polymorphism in the detected CpG sites are removed to obtain the methylation Beta value of the genomic CpG sites. The methylation level of the CpG site corresponding to the obtained gene of interest is extracted from the CpG site of the subject's genome, and the average value of the methylation values ​​of all CpG sites contained in each gene is defined as the methylation level of the gene.

[0074] Furthermore, the prediction module 30 is also used to obtain the symptom change data of the subject after treatment, use the symptom change data as the dependent variable of the preset linear regression equation, use the methylation level as the independent variable in the preset linear regression equation, and construct a stepwise multivariate linear regression model based on the dependent variable and the independent variable.

[0075] It is understandable that the symptom change data of the subject after treatment is obtained, for example, the symptom change of the patient after 8 weeks of treatment is used as the dependent variable, and the DNA methylation of the subject's gene of interest is used as the independent variable in the stepwise multiple linear regression equation, so as to construct a regression model for analysis.

[0076] Furthermore, the prediction module 30 is also used to evaluate the reduction rate of the symptom change data to obtain an evaluation result, perform a residual normality test on the multivariate linear regression model based on the evaluation result, and use the efficacy data that pass the test as the prediction result of the efficacy of the antipsychotic drug.

[0077] It should be understood that the PANSS score reduction rate can be used to evaluate the symptom changes of patients after 8 weeks of treatment, that is, (baseline score-post-treatment score) / (baseline score-30)×100%. The present invention uses the Kolmogorov-Smirnov test to test the normality of the dependent variable and the residual of the stepwise multiple linear regression model. In the stepwise multiple linear regression model, DNA methylation of genes of interest (such as HCN1 and MED19) that are significantly associated with symptom changes is detected, and the model containing DNA methylation of HCN1 and MED19 is defined as a model that can predict the efficacy of antipsychotic drugs. Detailed information of the model is shown in Table 2:

[0078] Table 2 Detailed information of stepwise multiple linear regression model

[0079]

[0080] Accordingly, the expression values ​​of the two genes HCN1 and MED19, the Z values ​​of each edge, and the names of the 21 edges in this model are shown in Table 3.

[0081] Table 3 Expression values ​​of the two genes HCN1 and MED19, the Z values ​​of each edge, and the names of the 21 edges in the stepwise multiple linear regression model

[0082]

[0083]

[0084] It should be noted that, in this article, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or system including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or system. In the absence of further restrictions, an element defined by the sentence "comprises a ..." does not exclude the existence of other identical elements in the process, method, article or system including the element.

[0085] The serial numbers of the above embodiments of the present invention are only for description and do not represent the advantages or disadvantages of the embodiments.

[0086] The above are only preferred embodiments of the present invention, and are not intended to limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A device for predicting the efficacy of antipsychotic drugs, characterized in that: The prediction device for the efficacy of antipsychotic drugs comprises: a variation determination module, for processing and calculating the volume of the subject's subcortical structure, calculating the correlation coefficient between the volumes of two brain regions in the volume, and determining the degree of variation of the covariant connection of the subcortical structure according to the correlation coefficient; A gene acquisition module, used to acquire the whole genome gene expression value of the autopsy brain tissue corresponding to the subcortical structure, and determine the gene of interest according to the whole genome gene expression value and the degree of variation; A prediction module, used to obtain the methylation level corresponding to the gene of interest, construct a stepwise multiple linear regression model according to the methylation level, and obtain the prediction result of the efficacy of the antipsychotic drug through the multiple linear regression model; Wherein, the prediction module is also used to analyze the methylation data of the whole genome DNA of the peripheral blood of the subject, and obtain the methylation level of the CpG dinucleotide site corresponding to the gene of interest from the methylation data; The prediction module is further used to obtain symptom change data of the subject after treatment, and to construct a stepwise multivariate linear regression model according to the methylation level and the symptom change data; The prediction module is also used to obtain the prediction result of the efficacy of antipsychotic drugs through the multivariate linear regression model; Among them, the prediction module is also used to obtain the symptom change data of the subject after treatment, use the symptom change data as the dependent variable of the preset linear regression equation, use the methylation level as the independent variable in the preset linear regression equation, and construct a stepwise multivariate linear regression model based on the dependent variable and the independent variable.

2. The device for predicting the efficacy of antipsychotic drugs according to claim 1, characterized in that: The variation determination module is further used to process and calculate the brain MRI scan data of the subject, and determine the volume of the subcortical structure based on the brain MRI scan data; The variation determination module is further used to calculate the correlation coefficient between the volumes of the two brain regions in the volume using the Pearson correlation analysis method, taking each pair of brain regions as an edge and using the correlation coefficient as the covariate connection strength of each edge; The variation determination module is further used to obtain the covariant connection strength difference of the covariant connection strength of each edge, and determine the variation degree of the covariant connection of the subcortical structure according to the covariant connection strength difference.

3. The device for predicting the efficacy of antipsychotic drugs according to claim 2, characterized in that: The variation determination module is also used to process and calculate the brain MRI scan data of the subject, perform head motion correction, skull stripping, coordinate transformation and subcortical white matter and dark gray matter volume structure segmentation on the MRI scan data, and obtain the volume of the subcortical structure.

4. The device for predicting the efficacy of antipsychotic drugs according to claim 2, characterized in that: The variation determination module is further used to obtain symptom relief data of the subject after treatment, and divide the covariate connection strength into a response group and a non-response group according to the symptom relief data; The variation determination module is also used to use a Z test to compare the difference in covariant connection strength of each edge in the response group and the non-response group, obtain the Z value of each edge, form a Z statistical graph based on the Z value, and determine the degree of variation of the covariant connection of the subcortical structure through the Z statistical graph.

5. The device for predicting the efficacy of antipsychotic drugs according to claim 1, characterized in that: The gene acquisition module is also used to obtain the whole genome gene expression value of the autopsy brain tissue corresponding to the subcortical structure in a preset database; The gene acquisition module is further used to calculate the correlation value between the whole genome gene expression value and the variation degree by using the least square method, and obtain the gene whose correlation value is greater than a preset correlation threshold as the target gene; The gene acquisition module is further used to take the target gene and the intersection gene in the preset risk gene set, and take the intersection gene as the gene of interest.

6. The device for predicting the efficacy of antipsychotic drugs according to claim 1, characterized in that: The gene acquisition module is also used to acquire original gene expression data of autopsy brain tissue corresponding to the subcortical structure in a preset database; The gene acquisition module is also used to verify the gene annotation of the original gene expression data through probes, filter out probes that do not exceed background noise, and select representative target probes; The original gene expression data is atlas-aligned by the target probe, and the expression value is normalized to obtain the whole genome gene expression value of the autopsy brain tissue.

7. The device for predicting the efficacy of antipsychotic drugs according to claim 1, characterized in that: The prediction module is also used to analyze the methylation data of the whole genome DNA of the peripheral blood of the subject, obtain the polymorphic sites in the CpG dinucleotide sites from the methylation data, eliminate the polymorphic sites, obtain the methylation risk index values ​​of the remaining CpG sites, and obtain the average value of the methylation values ​​corresponding to the methylation risk index values ​​from the gene of interest as the methylation level.

8. The device for predicting the efficacy of antipsychotic drugs according to claim 1, characterized in that: The prediction module is also used to evaluate the reduction rate of the symptom change data to obtain an evaluation result, perform a residual normality test on the multivariate linear regression model based on the evaluation result, and use the efficacy data that pass the test as the prediction result of the efficacy of the antipsychotic drug.

Images