基于生物信息学和机器学习的心肌梗死后心室重构关键基因的筛选

Screening of hub genes for ventricular remodeling post-myocardial infarction based on bioinformatics and machine learning

目的:利用生物信息学和机器学习筛选心肌梗死后心室重构(VRpMI)中的关键基因,并探索其对VRpMI的诊断价值。方法:从GEO数据库分别下载GSE132143中健康人和VRpMI患者心室组织的测序数据,猪心肌梗死后6个月梗死区和梗死远端区组织的测序数据,GSE775中小鼠心肌梗死后48 h、8周及正常小鼠心室组织的表达谱数据。基于健康人和VRpMI患者心室组织的测序数据,利用edgeR包和加权基因共表达网络分析筛选重要差异表达基因(DEG);通过LASSO算法和SVM-RFE算法筛选关键基因,并利用自身数据和猪、小鼠数据分析关键基因诊断VRpMI的价值;最后,对关键基因开展单基因的基因集富集分析。结果:共获得355个重要DEG,从中筛选出1个关键基因即神经元正五聚蛋白2(NPTX2)。NPTX2在自身数据和小鼠、猪验证数据中的AUC(95%CI)分别为0.996(0.984~1.000)、0.972(0.895~1.000)和0.963(0.882~1.000)。单基因的基因集富集分析显示,NPTX2富集于心肌收缩、鞘脂类代谢、细胞凋亡、谷胱甘肽代谢等19个信号通路。结论:NPTX2可能为诊断VRpMI的潜在生物标志物。

Aim:To screen hub genes for ventricular remodeling post-myocardial infarction(VRpMI)based on bioinformatics and machine learning and explore the diagnostic value of hub genes in VRpMI.Methods:The sequencing data of ventricular tissue from healthy individuals and VRpMI patients and from pigs at 6 months after myocardial infarction in GSE132143 dataset,as well as the expression profile data from left ventricular myocardial tissue of myocardially infarcted mice and normal mice at 48 hours and 8 weeks after myocardial infarction in the GSE775 dataset were downloaded from the GEO database.Based on sequencing data from ventricular tissue of healthy individuals and VRpMI patients,significant differentially expressed genes(DEG)were screened using edgeR package and weighted gene co-expression network analysis.The least absolute shrinkage and selection operator(LASSO)and support vector machine-recursive feature elimination(SVM-RFE)algorithms were used to screen hub genes,and the area under the receiver operating characteristic curve(AUC)for diagnosing VRpMI was calculated using the downloaded data.Gene set enrichment analysis(GSEA)of single gene was carried out on hub genes.Results:Totally 355 significant DEG were obtained and one hub gene(NPTX2)was screened out by LASSO and SVM-RFE.The AUC(95%CI)of NPTX2 were 0.996(0.984-1.000),0.972(0.895-1.000)and 0.963(0.882-1.000)in own data,validation data(mice,pigs),respectively.Single gene GSEA showed that the high expression of NPTX2 was enriched to 19 signaling pathways including myocardial contraction,sphingolipid metabolism,apoptosis,and glutathione metabolism,etc.Conclusion:NPTX2 might be a potential biomarker for diagnosing VRpMI.