【摘要】 目的基于脂代谢相关基因的生物信息学分析构建肝癌预后预测模型,并评价其对肝癌患者预后的预测能力。方法从癌症基因组图谱数据库(TCGA)和国际癌症基因组联合体数据库(ICGC)下载肝癌患者肿瘤组织和正常肝组织的mRNA转录组数据和临床资料(TCGA数据以欧美人种为主、ICGC数据仅为亚洲人种),从基因集合富集分析(GSEA)下载与脂质代谢相关的基因集。以TCGA数据库中的mRNA表达数据为输入文件,通过GSEA软件对基因集进行基因富集分析。采用perl软件和R软件从基因集中筛选出在肝癌中差异表达的脂代谢相关基因,采用Cox回归分析筛选出与肝癌预后相关的关键脂代谢基因,构建肝癌患者预后预测模型。计算TCGA、ICGC数据库中肝癌患者的肝癌预后预测模型风险评分,根据评分中位数将肝癌患者分为高风险者和低风险者,并进行Kaplan-Meier生存分析,绘制受试者工作曲线(ROC);根据TCGA数据库中患者的临床资料绘制诺莫列线图评价预后预测模型对肝癌预后的预测能力,通过Cox回归分析该模型风险评分与肝癌临床病理特征对肝癌患者预后的影响。结果脂质代谢相关基因集中不同基因共757个,筛选其中109个表达上调的基因进行单因素Cox分析,结果显示21个脂代谢基因与预后相关。多因素Cox分析筛选出6个能独立影响肝癌预后的关键脂代谢基因,分别为DAGLA、PCSK9、PIGU、FABP6、GLA、ESYT3。根据6个关键脂代谢基因构建肝癌预后预测模型,风险评分=DAGLA×0.161 878 597+PCSK9×0.014 967 28+PIGU×0.043 461 843+FABP6×0.078 505 362+GLA×0.019 799 065+ESYT3×0.278 063 77。TCGA、ICGC数据库Kaplan-Meier生存分析结果均显示,高风险肝癌患者的生存时间低于低风险者(P<0.01);两数据库ROC曲线下面积分别为为0.732、0.662;诺莫列线图显示该模型预测肝癌患者预后的C值为0.706;多因素Cox回归分析显示,该模型的风险评分与肝癌患者的TNM分期可作为独立预后影响因子(P<0.001)。结论成功构建了基于脂代谢相关基因的肝癌预后预测模型,该模型对肝癌患者预后的预测能力较好。更多还原

【Abstract】 Objective analysis of genes related to lipid metabolism,and to analyze its correlation with clinicopathological features.Methods The mRNA transcriptome data and clinical data of HCC tissues and normal liver tissues were downloaded from the Cancer Genome Atlas(TCGA)Database and the International Cancer Genome Consortium(ICGC)Database(the main ethnic groups in TCGA database are Europeans and Americans,and we only download the ICGC data from Asian ethnic groups).Gene sets-related to lipid metabolism were downloaded from gene set enrichment analysis(GSEA),and gene enrichment analysis was performed on the gene sets by GSEA software(version 4. 0. 1). The lipid metabolism-related genes differentially expressed in HCC were screened by Perl and R software from gene sets. Cox regression analysis was used to screen out the lipid metabolism genes related to the prognosis of HCC,and the prognosis prediction model was constructed. Risk scores of the models for HCC patients in TCGA and ICGC databases were calculated. Then HCC patients were divided into the high risk group and low risk group according to the median score. Kaplan-meier survival analysis was performed for the two groups,and receiver operating curve(ROC)was plotted. According to the clinical data of patients in TCGA database,nomogram was drawn to evaluate the predictive ability of the model for HCC patients prognosis. Cox regression was used to analyze the impact of the risk score and the clinical parameters on the prognosis of patients with HCC.Results were 757 lipid metabolism-related genes in the gene sets,among which,109 up-regulated genes were screened for univariate Cox analysis. The results showed that 21 lipid metabolism genes were associated with prognosis. Multivariate Cox analysis screened out 6 lipid metabolism genes that could independently affect the prognosis of HCC,including DAGLA,PCSK9,PIGU,FABP6,GLA,and ESYT3. Six lipid metabolism genes were used to construct a prognostic model for HCC.Risk score =DAGLA×0. 161 878 597+PCSK9×0. 014 967 28+PIGU×0. 043 461 843+FABP6×0. 078 505 362+GLA×0. 019 799 065+ESYT3×0. 27 806 377. Kaplan-Meier survival analysis results in TCGA and ICGC databases showed that the survival time of HCC patients in the high-risk group was lower than that in the low-risk group(P<0. 01). The area under ROC curves of the two databases was 0. 732 and 0. 662. The C-index of the nomogram was 0. 706. Multivariate Cox regression analysis showed that the risk score of the model and TNM stage of HCC patients could be used as independent prognostic factors(P<0. 001).Conclusion genes is successfully constructed,and the prediction ability of the model for the prognosis of HCC patients is good.更多还原