基于TCGA数据初步探查肝内胆管癌预后相关基因

目的 分析tCGA数据库中肝内胆管癌（ICC）高通量测序数据，寻找其预后相关基因，构建风险模型，并研究其在ICC组织中表达及作用通路。方法 下载tCGA数据库中33例ICC组织和8例癌旁组织中的RNA-seq表达矩阵数据和患者临床资料信息，利用edgeR软件包进行基因差异表达分析，通过单因素Cox回归分析筛选出预后相关差异基因，对差异基因绘制生存曲线，筛选出具有临床意义的基因，经多因素Cox回归分析并构建风险模型，通过京都基因与基因组百科全书（KEGG）通路富集分析了解预后相关基因的作用通路。结果 通过edgeR分析后得到6 617个差异基因（筛选标准为|log2 Fold Change|＞1，P＜0.05），其中高表达组4 094个，低表达组2 523个。通过功能富集发现，这些基因主要集中在化学物致癌作用、药物代谢-细胞色素P450系统、细胞色素P450对异生物质的代谢影响以及视黄醇代谢通路。经单因素Cox回归、R软件“survival”包生存曲线分析显示，UCN2、CST1、PROS1、SLC35E4、PEMT五个基因对ICC患者预后存在显著性影响。通过多因素Cox回归分析，CST1、PEMT、PROS1构建的风险模型对ICC患者预后具有判断作用。结论 UCN2、CST1、PROS1、SLC35E4、PEMT基因可能成为ICC预后判断指标，为后续临床试验提供数据支持。

Preliminary exploration of prognosis-related genes of intrahepatic cholangiocarcinoma based on TCGA database

Objective To find out the genes related to the prognosis of intrahepatic cholangiocarcinoma(icc), to build a risk model and to explore its expression and action pathway in the tissues of icc by analyzing the high-throughput sequencing data of icc in TCGA database. Methods RNA-seq expression matrix and clinical data of 33 cases of icc tissues and 8 cases of paracancerous tissues in TCGA database were downloaded. Gene expression data were analyzed by edge software package and the differential genes related to prognosis were selected by single factor Cox regression analysis. The survival curve was mapped and genes with clinical values were screened out. The risk model was built with multivariate Cox regression and its action pathway was reavealed through the enrichment analysis of Kyoto Encyclopedia of genes (KEGG). Results Under the screening criteria of“|log2 Fold Change|>1, P<0.05”, a total of 6 617 differentially expressed genes were obtained by edgeR analysis, including 4 094 in the high expression group and 2 523 in the low expression group. It was found that these genes were mainly concentrated in chemical carcinogenesis, drug -P450 system of exogenous substances metabolism, P450 on the metablism of xenobiotics and retinol metabolism through functional enrichment. Single factor Cox regression and R software“survival”package analysis showed that UCN2, CST1, PROS1, SLC35E4, PEMT had significant impacts on the prognosis of patients with icc. Multivariate Cox regression analysis showed that the risk models constructed by CST1, PEMT, and PROS1 could be used to predict the prognosis of patients with icc. Conclusion Gene UCN2, CST1, PROS1, SLC35E4, and PEMT may be the prognostic indicators for icc, which can provide data support for subsequent clinical trials.