Exploring pathogenesis of venous thromboembolism and screening biomarkers after integrating bioin formatics and machine learning methods
You Hongjun
Gou Qiling
Zhao Qianqian
Dong Mengya
Wang Yi
Abstract：Objective To study the pathogenesis of venous thromboembolism(VTE),and screen biomarkers.Methods The microarray data of whole blood gene expression profile was analyzed by using bioinformatic analyses in VTE patients and healthy controls(HC).The analyses included identification of differentially expressed gene(DEGs),enrichment analysis,and network analysis of protein-protein interactions(PPI).The key genes were screened by using regression model of least absolute shrinkage and selection operator(LASSO)and analysis of support vector machine recursive feature elimination(SVM-RFE).The expressions of key genes were validated by using external datasets,and receiver operator characteristic(ROC)curve was developed for reviewing diagnostic efficiency of the key genes.Results There were 59 DEGs screened through comparing expression profile data between VTE patients and HC.The results of enrichment analysis and PPI network analysis indicated that these genes were mainly involved in cellular response to starvation,oxidative phosphorylation,myeloid cell homeostasis,intrinsic apoptotic signaling pathway,inflammatory response and ribosome.There were 14 key genes screened including BCL2A1,CD3G,COX7A2,CSTA,EVI2A,FKBP1B,HBD,IFI27,LSP1,LY96,MS4A4A,PFDN5,RPS27L,and XK.The results of external dataset validation showed that 11 key genes were significantly up-regulated in VTE group,and they were BCL2A1,CD3G,COX7A2,CSTA,FKBP1B,HBD,IFI27,LY96,MS4A4A,PFDN5 and XK.The results of ROC curve analysis showed that these 11 key genes had higher diagnostic efficiency.Conclusion BCL2A1,CD3G,COX7A2,CSTA,FKBP1B,HBD,IFI27,LY96,MS4A4A,PFDN5,and XK can serve as biomarkers for VTE.
Keywords：Venous thromboembolismBioinformaticsMachine learning methodsDifferentially expressed genesBiomarkers
Publication Date：2025-07-20
Online Publishing Date：2025-09-09(First online date of this platform, not the publication date of the document)
Pages：5( 781-785 )