基于生物信息学方法筛选影响胶质母细胞瘤的诊断及治疗的关键基因

Screening of Key Genes Affecting the Diagnosis and Treatment of Glioblastoma Based on Bioinformatics

目的:通过生物信息学方法对胶质母细胞瘤公共数据库进行信息挖掘以筛选胶质母细胞瘤的关键差异表达基因,预测胶质母细胞瘤的可能潜在治疗靶点基因。方法:从基因表达数据库(GEO)中下载脑胶质母细胞瘤数据,通过在线分析软件GEO2R筛选正常组织与胶质母细胞瘤组织差异表达基因。通过Venn图取三个差异基因集交集。共筛选出632个共有差异基因。再通过STRING网站里的在线分析软件绘制PPI (protein-protein interaction network,蛋白互作网络),并导入Cytoscape软件利用MCODE插件筛选关键基因。通过R软件的Goplot包对差异表达基因进行GO和KEGG功能富集分析。再通过UCSC网站的癌症基因浏览器获得各关键基因在脑胶质母细胞瘤中的表达。通过GEPIA2网站绘制各关键基因的生存曲线。通过Oncomine网站获得关键预后基因KDELR2、CDKN3及RELN基因在不同级别胶质瘤中的表达。结果:从GEO数据库中下载GSE104291、GSE15824、GSE66354的数据集,通过GEO2R分析出差异表达基因,再通过Venn图共筛选出632个共有差异表达基因。通过STRING网站构建蛋白互作网络,并导入Cytoscape软件筛选出关键基因。获得差异表达基因的GO及KEGG功能富集分析图。通过GEPIA2网站获得关键基因的总生存率及无病生存率,结果发现KDELR2、CDKN3及RELN基因的表达差异与总生存率(OS)和无病生存率(RFS)相关性具有统计学意义(KDELR2、CDKN3和RELN这三个基因的总生存率生存曲线Logrank P值分别为−10;KDELR2、CDKN3和RELN这三个基因的无病生存率生存曲线Logrank P值分别为【0.001,【0.001以及0.00032)。KDELR2及CDKN3在正常组织中低表达,在胶质瘤组织表达与级别成正相关;RELN在在正常组织中高表达,在胶质瘤组织表达与级别成负相关。结论:通过本研究发现胶质瘤OS和RFS与KDELR2及CDKN3基因呈负相关,与RELN基因呈正相关。结合这些基因的表达差异,由此推测KDELR2基因及CDKN3基因可能是脑胶质母细胞瘤的发生及进展的癌基因,而RELN基因可能是一个抑癌基因。

Objective: To mine the information of glioblastoma public database by bioinformatics method to screen the key differentially expressed genes of glioblastoma and to predict the potential therapeutic target genes of glioblastoma. Methods: The glioblastoma data were downloaded from the Gene Expression Omnibus database (GEO), and the differentially expressed genes between normal and glioblastoma tissues were screened by online analysis software GEO2R. The intersection of three differential gene sets was obtained by Venn map. A total of 632 common differential genes were screened. Then draw PPI (protein-protein interaction network) through the online analysis software in STRING website, and then import the Cytoscape software to screen key genes using MCODE plug-ins. The functional enrichment of differentially expressed genes was analyzed by GO and KEGG by Goplot package of R software. Then the expression of key genes in glioblastoma was obtained by cancer gene browser on UCSC website. The survival curve of key genes was drawn by GEPIA2 website. The expressions of key prognostic genes KDELR2, CDKN3 and RELN in different grades of gliomas were obtained by Oncomine website. Results: The data sets of GSE104291, GSE15824 and GSE66354 were downloaded from GEO database, and the differentially expressed genes were analyzed by GEO2R. A total of 632 common differentially expressed genes were screened by Venn map. The PPI network was constructed by STRING website, and the key genes were screened by Cytoscape software. GO and KEGG functional enrichment analysis maps of differentially expressed genes were obtained. The overall survival rate and disease-free survival rate of key genes were obtained by GEPIA2 website. The results showed that the expression differences of KDELR2, CDKN3 and RELN genes were significantly correlated with overall survival rate (OS) and disease-free survival rate (RFS) (the Logrank P values of the overall survival curves of KDELR2, CDKN3 and RELN were −10, respectively;the Logrank P values of the disease-free survival curves of KDELR2, CDKN3 and RELN were