应用基因集富集分析和加权基因共表达网络分析挖掘支气管肺发育不良的关键基因  

Identification of key genes associated with bronchopulmonary dysplasia through gene set enrichment analysis and weighted gene co-expression network analysis

作　　者：胡诤贇 史建伟 张灵凤 张涛[1] 陈曦 孙春美[1] Hu Zhengyun;Shi Jianwei;Zhang Lingfeng;Zhang Tao;Chen Xi;Sun Chunmei(Department of Pediatrics,Shanghai Songjiang District Central Hospital,Shanghai 201600,China;Department of Neurosurgery,The Affiliated Brain Hospital of Nanjing Medical University,Nanjing 210024,China)

机构地区：[1]上海市松江区中心医院儿科,上海201600 [2]南京医科大学附属脑科医院神经外科,南京210024

出　　处：《中华生物医学工程杂志》2023年第4期361-367,共7页Chinese Journal of Biomedical Engineering

基　　金：国家自然科学基金(81773968)。

摘　　要：目的挖掘与支气管肺发育不良(BPD)密切相关的hub基因,为揭示BPD发病机制提供理论依据。方法从高通量基因表达(GEO)数据库获取BPD患儿的血mRNA数据集GSE32472、GSE108756、GSE189582基因表达矩阵,利用加权基因共表达网络分析(WGCNA)筛选出关键模块,通过GO、KEGG功能富集和GSEA分析获取关键模块基因参与的生物学过程。将GSE32472关键模块基因与GSE108756差异表达基因取交集后得出hub基因,再经受试者工作特征曲线(ROC)联合Logistic回归分析后得出与BPD关系密切的3个hub基因,验证这3个hub基因在外部数据集GSE189582中的诊断效能。结果WGCNA分析得出GSE32472中与BPD最密切相关的red模块共370个基因,GO分析集中在细胞因子介导信号通路、白细胞迁移及趋化、中性粒细胞激活、免疫受体等生物学功能。GSEA发现BPD组主要集中在免疫相关的分泌颗粒、三级颗粒、分泌囊泡的信号转导上。KEGG富集在病毒相关的细胞因子受体交互信号通路、雌激素信号通路、糖代谢相关的信号通路。red模块基因与GSE108756中201个差异表达基因取交集后的ROC结果显示3个hub基因(IL1R2、ADM、FPR2)对诊断BPD具有良好的特异度和敏感度,曲线下面积(AUC)均>0.7。IL1R2、ADM、FPR2对诊断BPD的OR值达到1.70、2.16、2.47(均P<0.05)。3个hub基因构建的逻辑回归模型在GSE189582中绘制出的ROC曲线AUC达到0.882。结论通过构建基因共表达调控网络筛选出3个hub基因,可能成为潜在的BPD基因生物标志物,为揭示BPD发病机制奠定一定的理论基础。Objective To identify hub genes closely associated with bronchopulmonary dysplasia(BPD)and provide a theoretical basis for understanding the pathogenesis of BPD.Methods High-throughput gene expression datasets(GSE32472,GSE108756,GSE189582)of blood mRNA from BPD children were obtained from GEO database.Weighted gene co-expression network analysis(WGCNA)was employed to identify key modules.GO and KEGG functional enrichment and GSEA analysis were used to understand the biological processes involved in key module genes.Hub genes were derived by intersecting the key module genes from GSE32472 with differentially expressed genes from GSE108756.Receiver operating characteristic(ROC)curve analysis in conjunction with Logistic regression was conducted to determine the three hub genes closely related to BPD.The diagnostic efficiency of these three hub genes was validated in an external dataset GSE189582.Results The WGCNA analysis showed that there were 370 genes in red module that were most closely related to BPD in GSE32472.GO analysis revealed biological functions related to cytokine-mediated signaling pathways,leukocyte migration and chemotaxis,neutrophil activation,immune receptor activity and so on.GSEA indicated enrichment of pathways related to immune-related secretory granules,tertiary granules,and exocytotic vesicles in the BPD group.KEGG enrichment showed pathways associated with virus-related cytokine-receptor interactions,estrogen signaling,and sugar metabolism.ROC analysis of the intersection of red module genes with 201 differentially expressed genes from GSE108756 identified three hub genes(IL1R2,ADM,FPR2)with good specificity and sensitivity for diagnosing BPD(all AUC>0.7).IL1R2,ADM,and FPR2 had odds ratios for BPD diagnosis of 1.70,2.16,and 2.47,respectively(all P<0.05).The Logistic regression model built with these three hub genes achieved an AUC of 0.882 in the GSE189582 dataset.Conclusion Three hub genes were screened out through the construction of gene co-expression regulatory network,which may be pot

关 键 词：基因集富集分析 加权基因共表达网络分析 支气管肺发育不良 hub基因 信号通路 ROC曲线 LOGISTIC模型 

分 类 号：R56[医药卫生—呼吸系统]