Method, equipment and device for identifying osteoarthritis gene features and storage medium
By combining the processing of database data and machine learning methods, the identification of gene features related to osteoarthritis is solved, and the problem of lack of large-scale genomic data analysis in the existing technology is solved, and a more accurate diagnosis of osteoarthritis is achieved.
Patent Information
- Application Number
- CN202411722975.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-11-28
- Publication Date
- 2025-04-18
AI Technical Summary
The lack of methods for large-scale analysis of genomic data in the prior art has led to great challenges in the early diagnosis and effective treatment of osteoarthritis.
By processing the data in the database, differential genes were obtained and gene modules with highly synergistic changes with osteoarthritis were identified. Combined with intersection operations and machine learning methods, candidate genes were determined and gene characteristics were determined.
More accurate identification of osteoarthritis gene characteristics is achieved, and gene characteristics related to osteoarthritis are provided, so that diagnosis can be carried out more accurately.
Smart Images
Figure CN119851912A_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the field of medical technology, and specifically relates to a method, equipment, device and storage medium for identifying genetic characteristics of osteoarthritis. Background Art
[0002] Osteoarthritis (OA) is a chronic joint disease that primarily affects articular cartilage and causes joint pain and dysfunction. With the increasing global aging population, the incidence of OA has increased year by year, becoming one of the leading causes of disability worldwide. Its pathogenesis is complex, involving multiple factors such as genetics, environment, and biomechanics, making early diagnosis and effective treatment a huge challenge. Therefore, it is particularly important to find biomarkers associated with OA. Machine learning technology is increasingly used in the biomedical field. By analyzing large-scale genomic data, meaningful patterns and relationships are extracted from complex biological data. MR is a method that uses genetic variation as an instrumental variable to infer causal relationships. Summary of the invention
[0003] The purpose of the present invention is to provide a method, device, apparatus and storage medium for identifying osteoarthritis gene characteristics, which solves the problem of the lack of large-scale analysis of genomic data in the existing technology.
[0004] In order to achieve the above object, the technical solution of the present invention is implemented as follows: a method for identifying osteoarthritis gene characteristics, comprising:
[0005] Step S1, processing the data in the database to obtain differentially expressed genes and identify gene modules with high synergistic changes with osteoarthritis;
[0006] Step S2, performing an intersection operation on the differentially expressed genes obtained in step S1 and the gene modules with a high degree of synergistic changes with osteoarthritis to obtain intersection genes;
[0007] Step S3, obtaining candidate genes by combining the intersection genes obtained in step S2 with an integrated machine learning method;
[0008] Step S4: determining gene features according to the candidate gene obtained in step S3 and the expression data of the candidate gene in the database.
[0009] In some embodiments, the step S1 is specifically:
[0010] Step S11, the data in the database includes a training set, the training set is de-batched and then differentially analyzed to obtain differential genes, and a heat map is drawn for the differential genes to obtain expression data;
[0011] Step S12, finding the gene modules highly correlated with the osteoarthritis in the expression data, using whether the training set is an osteoarthritis sample as the trait, using the genes in the expression data to construct a WGCNA network, and screening out the gene modules significantly correlated with the osteoarthritis.
[0012] In some embodiments, step S12 is specifically as follows:
[0013] Step S121, calculating the MAD value of the genes in the expression data in the sample, taking the first 5000 genes for WGCNA analysis, checking whether there are abnormal outlier samples in the sample, and performing sample hierarchical clustering on the data;
[0014] Step S122, calculating the soft threshold required for constructing the network, merging similar modules according to the soft threshold, and screening out gene modules with highly synergistic changes with osteoarthritis.
[0015] In some embodiments, the integrated machine learning method specifically integrates 12 machine learning algorithms to construct 113 models.
[0016] In some embodiments, step S3 is specifically the expression data of the candidate gene in the training set and the validation set, cross-validation is performed in the training set to obtain 113 models, the AUC value of each model in all validation sets is calculated, the final model is determined according to the AUC value, and the candidate gene is determined according to the final model.
[0017] In some embodiments, step S4 further comprises performing a correlation analysis on the expression of the gene signature in the training set.
[0018] In some embodiments, the machine learning algorithm is specifically a random forest, a least absolute shrinkage and selection operator, a ridge regression, an elastic net, a Stepglm, a support vector machine, component least squares estimation, a linear discriminant analysis, a gradient boosting machine, an extreme gradient boosting, and a naive Bayes classifier.
[0019] Another technical solution of the present invention is implemented as follows: a device for identifying osteoarthritis gene characteristics, comprising:
[0020] Data processing module: Process the data in the database to obtain differentially expressed genes and identify gene modules with high synergistic changes with osteoarthritis;
[0021] Intersection gene processing module: performing intersection operation on the differentially expressed genes and the gene module with highly synergistic changes with osteoarthritis to obtain intersection genes;
[0022] Candidate gene processing module: combining the intersection genes with the machine learning method to obtain candidate genes;
[0023] Gene feature processing module: determines gene features based on the candidate genes and expression data in the training set and the validation set.
[0024] Another technical solution of the present invention is implemented as follows: a device for identifying osteoarthritis gene characteristics, the device for identifying osteoarthritis gene characteristics: a memory, a processor, and a program for identifying osteoarthritis gene characteristics stored in the memory and executable on the processor, the program for identifying osteoarthritis gene characteristics being configured to implement the method for identifying osteoarthritis gene characteristics.
[0025] Another technical solution of the present invention is implemented as follows: a computer-readable storage medium storing a computer program, characterized in that the computer program can implement the steps in the method for identifying osteoarthritis gene characteristics when executed by a processor.
[0026] Compared with the prior art, the present invention more comprehensively analyzes and verifies the importance of characteristic genes in the pathogenesis of osteoarthritis, and provides clinicians with genetic features related to osteoarthritis, thereby enabling a more accurate diagnosis. BRIEF DESCRIPTION OF THE DRAWINGS
[0027] Figure 1 A schematic diagram of the structure of a device for identifying osteoarthritis gene characteristics provided by an embodiment of the present invention;
[0028] Figure 2 A flow chart of a method for identifying osteoarthritis gene characteristics provided by the first embodiment of the present invention;
[0029] Figure 3 A flow chart of a method for identifying osteoarthritis gene characteristics provided by a second embodiment of the present invention;
[0030] Figure 4 a is the PCA visualization of samples before batch removal in the training set;
[0031] Figure 4 b is the PCA visualization of the samples after the training set is batched;
[0032] Figure 5 This is the result diagram of difference analysis;
[0033] Figure 6 It is the sample hierarchical clustering diagram;
[0034] Figure 7 is the scale-free fitted exponential graph under soft thresholding power;
[0035] Figure 8 It is the model tree clustering diagram;
[0036] Fig. 9This is a combined result diagram of gene clustering and module partitioning;
[0037] Fig.10 This is the correlation analysis diagram between the module and osteoarthritis;
[0038] Fig.11 is the intersection venn graph;
[0039] Fig.12 This is the enrichment analysis result diagram;
[0040] Fig.13 a is the result of GO enrichment analysis;
[0041] Fig.13 b is the KEGG enrichment analysis result diagram;
[0042] Fig.14 a is the PPI network diagram constructed by STRING;
[0043] Fig.14 b is the MCODE key clustering network diagram;
[0044] Fig.15 Screening gene venn graphs for the cytoHubba algorithm;
[0045] Fig.16 To integrate the AUC results heatmap of 113 machine learning models;
[0046] Fig.17 a is the graph of λ and model characteristic coefficients;
[0047] Fig.17 b is the Enet cross-validation deviation change diagram;
[0048] Fig.17 c is the model characteristic coefficient diagram;
[0049] Fig.18 Diagnostic ROC curves for risk scores of each data set;
[0050] Fig.19 Diagnostic confusion matrix plots for risk scores of each data set;
[0051] Fig. 20 a is the box plot of characteristic gene expression in the training set;
[0052] Fig. 20 b is the ROC curve of characteristic genes;
[0053] Fig. 20 c is the Spearman correlation analysis diagram of characteristic genes;
[0054] Fig.21a is the box plot and ROC curve of the characteristic genes in the GSE89408 validation set;
[0055] Fig.21 b is the box plot and ROC curve of the characteristic genes of the GSE206848 validation set;
[0056] Fig.21 c is the box plot and ROC curve of the characteristic genes in the GSE236924 validation set;
[0057] Fig. 22 a is the heat map of immune cell abundance among OA diagnostic groups;
[0058] Fig. 22 b is the heat map of Spearman correlation analysis between immune cells;
[0059] Fig. 22 c is a box plot of immune cell abundance among OA diagnostic groups;
[0060] Fig.23 a is a bar graph showing the R value and P value of the correlation between immune cells and risk scores;
[0061] Fig.23 b Heat map of the correlation between characteristic genes and immune cells;
[0062] Fig.24 Predict characteristic genes and their related gene PPI network diagrams for the GeneMANIA database;
[0063] Fig.25 a is the GSEA pathway result diagram of all genes among GO enrichment risk score groups;
[0064] Fig.25 b is the GSEA pathway result diagram of all genes among KEGG enrichment risk score groups;
[0065] Fig.26 It is a bar chart showing the correlation between Hallmark pathway score and risk score;
[0066] Fig. 27 Results for nomogram construction and validation plots of training sets;
[0067] Fig.28 Box plot of sample nomogram scores between OA diagnostic groups in the validation set and ROC curve of nomogram prediction of diagnostic groups;
[0068] Fig.29 It is a mountain map of GSEA enrichment TOP10 for characteristic gene grouping;
[0069] Fig.30 a is the visualization result diagram of CTSO docking BX912;
[0070] Fig.30 b is the visualization result diagram of KIT docking with AVAPRITINIB;
[0071] Fig.30 c is the visualization result of PRDX4 docking with DOCETAXEL;
[0072] Fig.30 d is the visualization result diagram of PRKACB docking with FASUDIL;
[0073] Fig.30 e is the visualization result of PSMB9 docking with CARFILZOMIB;
[0074] Fig.30 f is the visualization result of TBL1XR1 docking ALLOPURINOL;
[0075] Fig.31 The forest plot is the MR analysis result;
[0076] Fig.32 a is a scatter plot of gene PSMB9 for SNPs in osteoarthritis MR analysis;
[0077] Fig.32 b is the forest plot of gene PSMB9 for SNPs in osteoarthritis MR analysis;
[0078] Fig.33 a is the leave-one-out analysis funnel plot;
[0079] Fig.33 b is the forest diagram of the validation results of leaving one SNP out one by one;
[0080] Fig.34 This is a structural block diagram of the first embodiment of the device for identifying osteoarthritis gene features of the present invention. DETAILED DESCRIPTION
[0081] 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.
[0082] Reference Figure 1 , Figure 1 This is a schematic diagram of the structure of a device for identifying osteoarthritis gene characteristics in the hardware operating environment involved in an embodiment of the present invention.
[0083] like Figure 1As shown, the device for identifying osteoarthritis gene features may include: a processor 1001, such as a central processing unit (CPU), a communication bus 1002, a user interface 1003, a network interface 1004, and a memory 1005. Among them, the communication bus 1002 is used to realize the connection and communication between these components. The user interface 1003 may include a display screen (Display), an input unit such as a keyboard (Keyboard), and the optional user interface 1003 may also include a standard wired interface and a wireless interface. The network interface 1004 may optionally include a standard wired interface and a wireless interface (such as a wireless fidelity (Wireless-Fidelity, Wi-Fi) interface). The memory 1005 may be a high-speed random access memory (Random Access Memory, RAM), or a stable non-volatile memory (Non-Volatile Memory, NVM), such as a disk storage. The memory 1005 may also be a storage device independent of the aforementioned processor 1001.
[0084] Those skilled in the art will understand that Figure 1 The structure shown in the figure does not constitute a limitation on the device for identifying osteoarthritis gene characteristics, and may include more or less components than shown in the figure, or combine certain components, or arrange the components differently.
[0085] like Figure 1 As shown, the memory 1005 as a storage medium may include an operating system, a network communication module, a user interface module, and a program for identifying osteoarthritis gene features.
[0086] exist Figure 1 In the device for identifying osteoarthritis gene characteristics shown, the network interface 1004 is mainly used for data communication with the network server; the user interface 1003 is mainly used for data interaction with the user; the processor 1001 and the memory 1005 in the device for identifying osteoarthritis gene characteristics of the present invention can be set in the device for identifying osteoarthritis gene characteristics, and the device for identifying osteoarthritis gene characteristics calls the osteoarthritis gene characteristics identification program stored in the memory 1005 through the processor 1001, and executes the osteoarthritis gene characteristics identification method provided by the embodiment of the present invention.
[0087] The present invention provides a method for identifying osteoarthritis gene characteristics, referring to Figure 2 , Figure 2 FIG. 4 is a flow chart of a first embodiment of a method for identifying osteoarthritis gene characteristics according to the present invention.
[0088] A method for identifying a genetic signature for osteoarthritis comprising:
[0089] Step S1, processing the data in the database to obtain differentially expressed genes and identifying gene modules with high synergistic changes with osteoarthritis;
[0090] Step S2, performing intersection operation on the differentially expressed genes in step S1 and the gene modules with highly synergistic changes with osteoarthritis to obtain intersection genes;
[0091] Step S3, obtaining candidate genes according to the intersection genes in step S2 in combination with an integrated machine learning method;
[0092] Step S4: determining gene features based on the candidate genes and the expression data in the training set and the validation set.
[0093] Reference Figure 3 , Figure 3 FIG. 4 is a flow chart of a second embodiment of a method for identifying osteoarthritis gene characteristics according to the present invention.
[0094] Based on the above first embodiment, in this embodiment, step S1 is specifically as follows:
[0095] Step S11, the data in the database includes a training set, the training set is de-batched and then differentially analyzed to obtain differential genes, and a heat map is drawn for the differential genes to obtain expression data;
[0096] Step S12, finding gene modules highly correlated with osteoarthritis in the expression data, using whether the training set is an osteoarthritis sample as a trait, using the genes in the expression data to construct a WGCNA network, and screening out gene modules significantly correlated with osteoarthritis.
[0097] It should be noted that in the specific implementation, all the data analyzed in this embodiment are from the NCBI GEO public database, as shown in Table 1.
[0098] Table 1 NCBI GEO dataset information table
[0099]
[0100]
[0101] For all GPL96 and GPL570 data sets, the data cel original files and the corresponding sequencing platform annotation files were downloaded from the GEO database, and then the R package was used to perform rma standardization on the data and extract the probe expression matrix. The annotation file was used to convert the probe ID to Gene Symbol, and the probe data without the corresponding Gene Symbol was removed. At the same time, for the case where multiple probes correspond to the same gene, the mean expression value of the corresponding different probes was taken as the final expression value of the gene. For the GPL11154 data, the R package lumi was used to process the data, and the standardization and the same gene symbol annotation were also performed. At the same time, the R package sva was used to merge the training set data and observe whether there was a batch effect. If so, the batch effect was removed.
[0102] Furthermore, the training set includes the data before and after merging and batch removal, and the data after merging and batch removal. PCA visualization is performed on the training set before and after batch removal, such as Figure 4 As shown in Figure 2, it is found that there is an obvious batch effect in the merged training set data. After SVA batch removal, there is no obvious batch effect in the samples.
[0103] Furthermore, we used the limma package to perform a differential analysis of Disease vs Control, and finally obtained 408 differentially expressed genes, including 274 up-regulated genes and 134 down-regulated genes. The volcano plot of the differential analysis results is shown below: Figure 5 As shown in a, a heat map was drawn for the up-regulated and down-regulated TOP10 gene expressions, as shown in Figure 5 As shown in b.
[0104] Reference Figure 6 , Figure 6 FIG. 4 is a flow chart of a third embodiment of a method for identifying genetic features of osteoarthritis according to the present invention.
[0105] Based on the above second embodiment, in this embodiment, the step S12 is specifically as follows:
[0106] Step S121, calculate the MAD value of the gene in the sample, take the first 5000 genes for WGCNA analysis, check whether there are abnormal outliers in the sample, and perform sample hierarchical clustering on the data;
[0107] Step S122, calculating the soft threshold required for constructing the network, merging similar modules according to the soft threshold, and screening out gene modules with highly synergistic changes with osteoarthritis.
[0108] It should be noted that in the specific implementation, in order to find the gene modules that are highly correlated with osteoarthritis in the expression data, we used whether the sample is an osteoarthritis sample as the trait, and used the genes in the expression data to construct the WGCNA network to screen out the gene modules and genes that are significantly correlated with osteoarthritis.
[0109] First, the MAD value of the gene in the sample is calculated, and the first 5000 genes are taken for WGCNA analysis. Then, whether there are abnormal outliers in the sample is checked, and the data is clustered by sample level. The results are as follows: Figure 7 The results show that there are no obvious outlier samples, and subsequent analysis can be carried out normally.
[0110] Then we need to start calculating the soft threshold required to build the network. The horizontal axis of the figure below represents the weight parameter. The vertical axis of the left figure is Scale Free Topology Model Fit, which is signed R^2, which is the square of the correlation coefficient between log(k) and log(p(k)) in the corresponding network. The higher the square of the correlation coefficient, the closer the network is to the distribution without network scale. The vertical axis of the right figure represents the mean of all gene adjacency functions in the corresponding gene module. According to Figure 8 As shown in a, the soft threshold corresponding to the first time R^2 is greater than 0.85 is selected, which is 10, indicating that the network is close to scale-free distribution under this soft threshold, and Figure 8 The mean value of the number of gene connections in b gradually becomes flat, which conforms to the law.
[0111] After obtaining the appropriate soft threshold, it is necessary to establish an adjacency matrix based on the soft threshold, make the gene distribution conform to the scale-free network according to the connectivity, and then calculate the gene topology matrix based on the adjacency matrix, and then deduce the dissimilarity coefficient between genes, and then obtain the systematic clustering tree between genes. According to the hybrid dynamic tree cutting algorithm, the module division is performed, and the results are as follows Fig. 9 As shown in , similar modules are then merged according to the module tree clustering height, and the minimum number of genes in each gene module is set to 50. Finally, 12 modules are obtained except the gray module. The results are shown in Fig.10 shown.
[0112] After dividing the modules, the modules were analyzed for association with whether the samples were osteoarthritis. The results are as follows: Fig.11 shown.
[0113] According to the method, the intersection of differentially expressed genes (DEGs) and WGCNA key module genes was performed and the intersection genes were regarded as disease progression genes. The results are shown in Fig.12 As shown, 80 intersection genes were finally obtained.
[0114] Subsequently, enrichment analysis was performed on these 80 intersection genes, and the results with pvalue < 0.05 were screened as significant enrichment results. Finally, 107 GO BP terms, 5 GO CC terms, 10 GO MF terms and 12 KEGG pathways were obtained, and the results were visualized, as shown in the figure. Fig.13 as shown.
[0115] The STRING database was used to construct a PPI network for the common differentially expressed genes, and genes without network relationships were removed. The PPI network structure was visualized using Cytoscape, and a network structure containing 77 gene nodes and 180 association relationships was obtained. Fig.14 As shown in a.
[0116] Then, MCODE was used to extract key clusters from the PPI network, and a cluster network containing 67 gene nodes and 171 association relationships was obtained. Fig.14 As shown in b, cytoHubba was used to extract key genes using four algorithms, and the TOP 30 genes were extracted respectively. The intersection operation was performed on the four gene sets. The results are shown in Fig.15 As shown, 20 key candidate genes were finally obtained.
[0117] Based on the expression data of the above 20 key candidate genes in the training set and multiple validation sets, 10-fold cross validation was performed in the training set to establish the 113 best-performing final models. The AUC (area under the ROC curve and the coordinate axis) value of each model in all validation sets was further calculated, and a heat map was drawn based on the AUC results of all models, as shown in the figure below. Fig.16 As shown. The final elastic network Enet model achieved the best performance and the highest average AUC result (0.871). Its alpha coefficient was set to 0.8. The model finally selected 12 genes, and its model characteristic risk score = ARMCX2*0.7507+PRDX4*2.0179+MYL6B*1.4038+
[0118] RRAS*1.8518+STMN2*2.0221+CTSO*1.3655+TBL1XR1*0.4694+RARRES3*0.9553-KIT*0.4976+PRKACB*0.7098+ZIC1*2.1770+PSMB9*1.4912, the final elastic network optimal model tuning process and coefficient selection are as follows Fig.17 shown.
[0119] After obtaining the optimal model, the logistic model constructed with 12 model characteristic genes was used to calculate the risk score of each patient, and the sample diagnosis ROC curve was drawn in the training set and three validation data sets according to the risk score. The results are shown in Fig.18 As shown. And calculate the sample diagnosis prediction confusion matrix, as Fig.19 shown.
[0120] The differential expression of the characteristic genes of the above model was verified using the training set and validation set data. The expression level box plots of the 12 characteristic genes obtained from the above analysis in the validation set and all training sets were drawn, and the single gene ROC curves of the characteristic genes were drawn. The training set results are shown in Fig. 20 The validation set results are shown in Fig.21 As shown. Spearman correlation analysis was performed on the expression of all characteristic genes in the training set.
[0121] The training set was subjected to ssGSEA immune infiltration analysis to obtain immune cell infiltration data. The immune cell infiltration data and expression data were combined to observe the potential association between immune cells and diagnostic groups, draw heat maps and box plots of differences between immune cells and groups, and calculate the correlation heat maps between immune cells, such as Fig. 22 shown.
[0122] At the same time, in order to observe the association between the model characteristic genes and the model risk score and immune cells, Spearman correlation analysis was performed on immune cells, risk scores and characteristic genes. Fig.23 shown.
[0123] Based on the GeneMANIA database, PPI network analysis was performed using characteristic genes, and PPI networks were constructed using characteristic genes and their 20 interacting genes to predict colocalization, shared protein domains, coexpression, prediction, and correlation between pathways. The results are shown in Fig.24 shown.
[0124] We then divided the OA samples into high-risk and low-risk groups according to the median risk score, calculated the logFC values of genes in the high-risk and low-risk groups, sorted them from large to small, and then performed GSEA analysis (gene enrichment analysis). Pathways with pvalue < 0.05 & |NES| > 1 were selected as valid results. The results are visualized as follows: Fig.25 At the same time, the GSVA algorithm was used to perform enrichment scoring on OA samples based on the hallmark gene set, and the correlation between the risk score and the hallmark enrichment score was calculated, and a correlation bar chart was drawn as shown in Fig.26 shown.
[0125] According to the method, based on the above-mentioned characteristic genes, a nomogram was constructed to achieve clinical reference significance. At the same time, in order to verify the reliability of the nomogram, calibration curves, decision curves and ROC curves were drawn to evaluate it. The results are shown in Fig. 27 As shown, diagnosis prediction was also performed in the validation set, and the ROC curve was drawn. A box plot of the difference in diagnostic scores between the true diagnostic groups was drawn according to the diagnostic scores. The results are shown in Fig.28 shown.
[0126] According to the nomogram, the patient's diagnostic score can be calculated intuitively, and then the patient's diagnostic results can be predicted. At the same time, in order to verify the accuracy of the nomogram, a calibration curve and a decision curve are also drawn. The solid line in the calibration curve represents our prediction model, and the dotted diagonal line represents the actual situation. It can be found that the prediction of our model is close to the actual situation, and it has good prediction performance overall; and the decision curve can show the net benefit of different decision schemes within a specific threshold range. In the range of 0-1, if the model is higher than the All line and the None line, it means that the model has good actual clinical prediction performance. It can be seen from the figure that the nomogram model combined with the risk score has good actual clinical prediction performance.
[0127] According to the method, all characteristic genes were grouped into high and low expression groups, all genes were ranked according to the logFC difference of single gene grouping, GSEA analysis was performed based on the KEGG background gene set, and a mountain map of the TOP 10 pathways was drawn for display.
[0128] Based on the small molecule drug correspondence obtained by searching the DGidb database based on 12 characteristic genes, 6 genes and their drug associations were finally obtained. Then, the PDB format data of the target protein corresponding complexes of these six genes were found in the PDB database, and the 3D structure SDF file of the active ingredient was found in the PubChem database. Pymol was used to pre-process it, remove small molecule structures outside the target protein, dehydrate and hydrogenate, and retain the target protein structure. Autodocktool was used to set the protein gridbox coverage area to include the entire protein receptor, and then the 6 proteins were molecularly docked and verified in Autodockvina, and their docking scores were calculated. Then, the results with the smallest docking score and connection bonds for each active ingredient were selected for visualization. Some results are shown below. Fig.29 shown.
[0129] As described in the methods, Mendelian randomization analysis was performed using the R package, with eQTL and mQTL data as exposure data and osteoarthritis as outcome data. All valid eQTL and pQTL data are listed in gene.exposure.csv and protein.exposure.csv. The SNP information of characteristic genes in eQTL and pQTL was screened out, and MR analysis was performed in turn with multiple osteoarthritis data. Results with IVW fixed effects or IVW random effects with Pval less than 0.05 were retained.
[0130] The ebi-a-GCST007090 data are presented here as the results of the outcome.
[0131] After analysis, the MR results of four genes, PSMB9, RPAS, CTSO, and RARRES3 (PLAAT4), were retained in the final eQTL data. The forest plot of the MR analysis results is shown in Figure 2. Fig.30 shown.
[0132] The results show that the IVW fixed effect of PSMB9 shows a risk-type causal relationship for osteoarthritis. Therefore, the MR analysis process diagram of the PSMB9 gene is shown as follows: Fig.31 shown.
[0133] At the same time, sensitivity analysis and pleiotropy level test were performed on the results of PSMB9. The results are shown in Tables 2 and 3 below. It was found that there was no heterogeneity and pleiotropy in the results, proving that the results are reliable.
[0134] Table 2
[0135]
[0136] Table 3
[0137] exposure outcome egger_intercept SE p-value PSMB9 Osteoarthritis -0.018 0.011 0.160
[0138] The final results showed that the gene PSMB9 showed a risk-type causal relationship with osteoarthritis, and there was no heterogeneity and pleiotropy.
[0139] refer to Fig.34 , Fig.34 This is a structural block diagram of a first embodiment of a device for identifying osteoarthritis gene characteristics according to the present invention.
[0140] like Fig.34 As shown, the device for identifying the genetic signature of osteoarthritis comprises:
[0141] Data processing module: Process the data in the database to obtain differentially expressed genes and identify gene modules with high synergistic changes with osteoarthritis;
[0142] Intersection gene processing module: performing intersection operation on the differentially expressed genes and the gene module with highly synergistic changes with osteoarthritis to obtain intersection genes;
[0143] Candidate gene processing module: combining the intersection genes with the machine learning method to obtain candidate genes;
[0144] Gene feature processing module: determines gene features based on the candidate genes and expression data in the training set and the validation set.
[0145] It should be understood that the above is only an example and does not constitute any limitation on the technical solution of the present invention. In specific applications, technicians in this field can make settings as needed, and the present invention does not limit this.
[0146] It should be noted that the workflow described above is merely illustrative and does not limit the scope of protection of the present invention. In practical applications, technicians in this field can select part or all of them according to actual needs to achieve the purpose of the present embodiment, and no limitation is made here.
[0147] In addition, for technical details not fully described in this embodiment, reference can be made to the method for identifying osteoarthritis characteristic genes provided in any embodiment of the present invention, and will not be repeated here.
[0148] In addition, 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.
[0149] 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.
[0150] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as a read-only memory (ROM) / RAM, a magnetic disk, or an optical disk), and includes a number of instructions for a terminal device (which can be a mobile phone, a computer, a server, or a network device, etc.) to execute the methods described in each embodiment of the present invention.
[0151] 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 method for identifying genetic characteristics of osteoarthritis, characterized in that: include: Step S1, processing the data in the database to obtain differentially expressed genes and identify gene modules with high synergistic changes with osteoarthritis; Step S2, performing an intersection operation on the differentially expressed genes obtained in step S1 and the gene modules with a high degree of synergistic changes with osteoarthritis to obtain intersection genes; Step S3, combining the intersection genes obtained in step S2 with the integrated machine learning method to obtain candidate genes; Step S4: determining gene features according to the candidate gene obtained in step S3 and the expression data of the candidate gene in the database.
2. The method for identifying osteoarthritis gene signature according to claim 1, characterized in that: The step S1 is specifically as follows: Step S11, the data in the database includes a training set, the training set is de-batched and then differentially analyzed to obtain differential genes, and a heat map is drawn for the differential genes to obtain expression data; Step S12, finding the gene modules highly correlated with the osteoarthritis in the expression data, using whether the training set is an osteoarthritis sample as the trait, using the genes in the expression data to construct a WGCNA network, and screening out the gene modules significantly correlated with the osteoarthritis.
3. The method for identifying osteoarthritis gene signature according to claim 2, characterized in that: The step S12 is specifically as follows: Step S121, calculating the MAD value of the genes in the expression data in the sample, taking the first 5000 genes for WGCNA analysis, checking whether there are abnormal outlier samples in the sample, and performing sample hierarchical clustering on the data; Step S122, calculating the soft threshold required for constructing the network, merging similar modules according to the soft threshold, and screening out gene modules with highly synergistic changes with osteoarthritis.
4. The method for identifying osteoarthritis gene signature according to any one of claims 1 to 3, characterized in that: The integrated machine learning method specifically integrates 12 machine learning algorithms to construct 113 models.
5. The method for identifying osteoarthritis gene signature according to claim 4, characterized in that: The step S3 is specifically the expression data of the candidate gene in the training set and the validation set, cross-validation is performed in the training set to obtain 113 models, the AUC value of each model in all validation sets is calculated, the final model is determined according to the AUC value, and the candidate gene is determined according to the final model.
6. The method for identifying osteoarthritis gene signature according to claim 5, characterized in that: The step S4 also includes performing a correlation analysis on the expression of the gene feature in the training set.
7. The method for identifying osteoarthritis gene signature according to claim 4, characterized in that: The machine learning algorithms are specifically random forest, least absolute shrinkage and selection operator, ridge regression, elastic net, Stepglm, support vector machine, component least squares estimation, linear discriminant analysis, gradient boosting machine, extreme gradient boosting and naive Bayes classifier.
8. A device for identifying genetic characteristics of osteoarthritis, characterized in that: include: Data processing module: Process the data in the database to obtain differentially expressed genes and identify gene modules with high synergistic changes with osteoarthritis; Intersection gene processing module: performing intersection operation on the differentially expressed genes and the gene module with highly synergistic changes with osteoarthritis to obtain intersection genes; Candidate gene processing module: combining the intersection genes with the machine learning method to obtain candidate genes; Gene feature processing module: determines gene features based on the candidate genes and expression data in the training set and the validation set.
9. A device for identifying genetic characteristics of osteoarthritis, characterized in that: The device for identifying osteoarthritis gene characteristics: a memory, a processor, and a program for identifying osteoarthritis gene characteristics stored in the memory and executable on the processor, wherein the program for identifying osteoarthritis gene characteristics is configured to implement the method for identifying osteoarthritis gene characteristics according to any one of claims 1 to 7.
10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps in the method for identifying osteoarthritis gene features according to any one of claims 1 to 7 can be implemented.