基于TCGA数据库的肝癌发生关键基因筛选

Identification Key Genes of Hepatocellular Carcinoma Base on TCGA Database

摘要: 目的：肝癌是消化系统常见恶性肿瘤，是全世界第三位死亡原因和中国第二位死亡原因。肿瘤基因组图谱(TCGA)计划利用大规模测序为主的基因组分析技术，通过广泛的合作，理解癌症的分子机制，本研究利用TCGA数据库深入挖掘肝癌发生关键基因。材料方法：根据标准流程对TCGA数据进行处理、整合，对数据类型及水平进行评估，用R语言(3.1.1版本)中自带的DESeq和edgeR程序包进行分析，结果以热图(pheatmap)、韦恩图(VennDiagram)、hist、PlotMA等表示。差异基因的判断标准：1，表达量在2倍以上或者0.5倍以下，2，P < 0.05，3，基因排名在前10%。结果：TCGA数据库现有癌组织mRNA芯片信息17张，匹配正常组织mRNA芯片信息9张，共26张。Hist图反映的是每个经统计后P值得分布规律，图中可刊出P值接近0处频率很高，反映差异基因的数量很大。PLotMA图反应基因表达量的分布规律，提示表达上升基因数量较多。用DESeq方法一共找到719个差异基因，而用edgeR方法找到4413个差异基因，两种方法都鉴别出的共同差异基因713个。结论：TCGA法相较于传统的芯片筛选具有样本数量大、费用小、分析简单等优势，为更多的人进行大规模的肝癌基因组学研究以及基于基因组学的后续功能研究提供了可能性。

Abstract: Objective: Hepatocellular carcinoma (HCC) is a common cancer of the digestive system, is the third cause of death worldwide and the second cause of death in China. The Cancer Genome Atlas (TCGA) aims to better understand the molecular mechanisms of cancer by using a large-scale genome se-quencing-based analysis techniques and extensive cooperation. This study introduces TCGA data-base to find key genes of HCC events. Materials and Methods: The data from TCGA were processed, integrated according to the standard procedure of TCGA, data types and levels were carefully as-sessed. Bioinformatics analysis was done using the DESeq and edgeR package of R language (3.1.1 version). Results were showed as pheatmap, VennDiagram, hist, PlotMA etc. Differences were de-fined as follows: expression increased more than two folds; P <0.05; gene ranked in the top 10%. Results: 17 mRNA chips of HCC and 9 mRNA chips of normal tissue were collected from TCGA data- base. Hist figure reflected the number of different gene was large. PLotMA map showed the distribu-tion of gene expression, suggesting most genes of different expression were increased. 719 diffe-rentially expressed genes were found by DESeq, while 4413 by edger, among which 713 were com-mon different genes. Conclusion: Compared to conventional microarray, TCGA method has its own advantages such as larger number of samples, less cost and easier for analyzing, offering opportuni-ty for large-scale genomic studies of HCC and subsequent functional genomics-based research.