Exploring signature gene and predicting the target active ingredients of TCM for immunosuppression in sepsis based on weighted gene co-expression network analysis and machine learning
Objective To identify immunosuppression in sepsis signature genes and predict the active ingredient of traditional Chinese medicine(TCM)using weighted gene co-expression network analysis(WGCNA)and variable machine learning methods.Method The GSE182522 dataset was downloaded from the gene expression omnibus,and differentially expressed genes(DEGs)were extracted using the limma software package of Bioconductor.Kyoto Encyclopedia of Genes and Genomes(KEGG)pathway and gene ontology(GO)enrichment analysis were performed on DEGs.Additionally,WGCNA and the least absolute shrinkage and selection operator(LASSO)regression,support vector machine-recursive feature elimination(SVM-RFE),and elastic network regression were used to screen immunosuppression in sepsis signature genes.Then,the receiver operating characteristic curve(ROC curve)was used to evaluate the diagnostic performance,and draw a box plot to show the expression patterns of the feature genes.A total of 39 SPF BALB/c mice were randomly assigned to 2 groups:immunocompetent-sepsis group(n=20)and immunosuppressed-sepsis group(n=19).The animal model of sepsis was established by cecal ligation and perforation.Peripheral blood of mice in immunocompetent-sepsis group or immunosuppressed-sepsis group were collected at 12 h or 24 h,respectively.Reverse transcription quantitative polymerase chain reaction(RT-qPCR)was used to detect the expression of TRBV7-2 in two groups of mice,and to verify the diagnostic efficacy of TRBV7-2.Finally,the potential active ingredient of TCM was predicted by molecular docking technology.Result A total of 445 DEGs were screened out,among which 173 genes were upregulated and 272 genes were downregulated.Functional enrichment analysis revealed that these DEGs were mainly involved in three signaling pathways:nitrogen metabolism,bile secretion and gastric acid secretion.Additionally,these DEGs were significantly enriched in various biological processes,including the positive regulation of molecular function,cellular response to oxygen-containing compounds,cellular response to nitrogen-containing compounds,cellular response to peptide hormone stimuli,and the regulation of glucose import.In terms of cellular components,the DEGs were notably linked to the plasma membrane protein complex,T-cell receptor complex,and lateral plasma membrane.At the molecular function level,these genes were particularly involved in nucleoside triphosphatase regulator activity,GTPase activator activity,and exogenous protein binding.A total of 56 key genes were screened by WGCNA,and one signature gene was obtained by the intersection of 3 kinds of machine learning:TRBV7-2.ROC curve analysis showed that TRBV7-2 had high clinical diagnostic value and was significantly down-regulated in patients with immunosuppression in sepsis samples.The results of RT-qPCR showed that the expression of TRBV7-2 gene in blood of immunosuppressed-sepsis group was lower than that of immunocompetent-sepsis group.By molecular docking technique,10 potential active ingredients of TCM,such as berberine,baicalin and atractylodin were predicted.Conclusion TRBV7-2 may be used as a diagnostic biomarker of immunosuppression in sepsis,and berberine and the 10 other active ingredients found in TCM may be provide a foundation for reducing the mortality of sepsis.
SepsisImmunosuppressionWeighted gene co-expression network analysisMachine learningSignature geneTargeted traditional Chinese medicine
万方数据
维普
