基于TCGA在宫颈癌肿瘤微环境中挖掘免疫疗效相关的预后基因

Exploration on prognosis genes related to immunization efficacy based on TCGA in cervical cancer tumor microenvironment

目的结合生物信息学方法,使用肿瘤基因组图谱(TCGA)数据库中宫颈癌表达谱数据,通过ESTIMATE算法分析肿瘤微环境挖掘有价值的预测免疫治疗疗效相关的预后基因。方法从TCGA数据库下载宫颈癌数据,运用R软件ESTIMATE包将宫颈癌基因表达谱数据分为高低免疫/高低基质评分组,通过limma包计算各组差异基因,最终两组的交集基因将用于生存分析、功能富集分析及蛋白互作网络构建。使用R Survival软件包分析高低免疫/高低基质评分组与预后的关系,同时使用Wilcox检验分析高低免疫/高低基质评分组与临床特征之间的关系。结果高免疫评分组与预后呈正相关(P=0.035);高基质评分组中位总生存期延长,但无统计学意义(P=0.391)。高免疫评分组、高基质评分组与M分期呈正相关(P<0.05)。设定阈值为|log2(FC)|>2且FDR<0.05,共鉴定出485个交集差异基因(上调475个,下调10个)。差异基因生存分析显示157个基因与预后有关。基因本体分析(GO)和京都基因与基因组百科全书途径(KEGG)共富集788个基因本体术语和36条通路(FDR<0.05)。差异基因构建PPI网络并利用Cytoscape提取3个关键核心网络,共包括66个基因,其中12个也与预后相关。结论基于TCGA数据库分析肿瘤微环境获得的免疫预后相关基因可能是潜在的免疫治疗疗效预测的生物标志物,值得深入研究。

Objective To combine the bioinformatics method and use the Cancer Genome Atlas (TCGA) database cervical cancer expression profile data to analyze the tumor microenvironment by using the ESTIMATE algorithm to explore valuable prognostic genes related to the prediction of immunotherapy efficacy. Methods Cervical cancer data were downloaded from the TCGA database. The cervical cancer gene expression profile data were divided into high and low immune/high and low matrix score groups by R software ESTIMATE package. The differential genes were calculated by limma package. And the final two sets of intersection genes would be used in survival analysis,functional enrichment analysis,and protein interaction network construction. R Survival package was used to analyze the relationship between the high and low immune/high and low matrix score groups and the prognosis,and the relationship between the high and low immune/high and low matrix score groups and clinical features was analyzed by Wilcox test. Results The high immune score group was positively correlated with prognosis (P=0.035);the median overall survival was prolonged in the high matrix score group,but difference was not statistically significant (P=0.391). The high immune score group and the high matrix score group were positively correlated with M stage (P<0.05). The threshold was set to | log2 (FC) | >2 and FDR<0.05. A total of 485 intersection difference genes were identified (475 up-regulation and 10 down-regulation). Differential gene survival analysis showed that 157 genes were associated with prognosis. The gene ontology (GO) analysis and the Kyoto Encyclopedia of Genes and Genome Pathways (KEGG) enriched 788 gene ontology terms and 36 pathways (FDR<0.05). The differential gene constructed PPI network and used Cytoscape to extract 3 key core networks,including 66 genes,12 of which were also associated with prognosis.Conclusion Immune prognosis-related genes obtained from tumor microenvironment based on TCGA database may be a potential biomarker for predicting the efficacy of immunotherapy,which is worthy of further study.