基于生物信息学途径食管鳞癌lncRNA预后风险模型的构建  

Establishment of a Prognostic Risk Model of lncRNA in Esophageal Squamous Cell Carcinoma by Bioinformatics

作　　者：蔡河源 谢春莹 邹健勇[1] 罗红鹤[1] CAI He-yuan;XIE Chun-ying;ZOU Jian-yong;LUO Hong-he(Department of Thoracic Surgery,The First Affiliated Hospital of Sun Yat-sen University,Guangzhou 510080,China)

机构地区：[1]中山大学附属第一医院胸外科,广东广州510080

出　　处：《中山大学学报（医学科学版）》2021年第5期729-737,共9页Journal of Sun Yat-Sen University:Medical Sciences

基　　金：卫生部医药卫生科技发展中心项目(W2014RQ22)。

摘　　要：【目的】基于癌症基因组图谱(TCGA)数据库食管鳞癌lncRNA差异表达,构建食管鳞癌预后风险模型。【方法】从TCGA肿瘤数据库下载食管鳞癌基因表达数据及临床信息,使用R 4.0.3软件提取差异表达lncRNA,进一步使用COX回归分析筛选模型lncRNA,并构建预后风险模型Riskscore;按照风险评分中位值将食管鳞癌患者分为低风险组和高风险组,进一步比较高低风险组生存预后;分析比较风险模型及其他临床特征对食管鳞癌生存预后的预测性能;分析风险评分与其他临床特征相关性;采用主成分分析(PCA)及基因富集分析(GSEA)探索高低风险组基因分布差异。【结果】得到差异表达lncRNA174个,食管鳞癌组织表达上调126个,表达下调48个。COX回归分析显示AL033384.1,AC108449.2可作为食管鳞癌的独立预后因素,风险模型公式为Riskscore=1.303×AL033384.1-1.525×AC108449.2,且低风险组总体生存率高于高风险组(P<0.001),主成分分析显示模型lncRNA能较好区分高低风险组,是高低风险组分布的主要影响因素。高低风险组存在细胞通路差异,高风险组主要富集在表皮细胞的角质化和细胞内的高代谢等过程。该风险模型1、2和3年预测性能分别为0.750、0.768和0.796,优于其他临床特征如TNM分期的预测性能。【结论】基于TCGA数据库筛选得到的AL033384.1,AC108449.2建立的预后风险模型对预测食管鳞癌患者的预后具有一定的临床价值,低风险组患者总体生存率更长,预后更好。【Objective】To establish a prognostic risk-model of lncRNA in esophageal squamous cell carcinoma based on cancer genome atlas(TCGA)database.【Methods】The gene expression data and clinical information of esophageal squamous cell carcinoma were downloaded from the TCGA database.The R 4.0.3 software was used to identify the differential expression of lncRNA,Cox regression analysis was used to screen the lncRNA of model,and a prognostic risk model-Riskscore was established.According to the median of Riskscore,the patients were divided into high and low risk group,and the survival prognosis between the two groups were further compared.The diagnostic performance of Riskscore and other clinical features of survival prognosis were further analyzed.In addition,the correlation between Riskscore and other clinical features was analyzed.Finally,principal component analysis(PCA)and gene enrichment analysis(GSEA)were used to explore the differences in gene distribution between the two groups.【Results】Using R 4.0.3 software,174 lncRNA were differentially expressed,in which 126 were up-regulated and 48 were down-regulated.Univariate and multivariate COX regression analysis showed AL 033384.1 and AC 108449.2 could be independent prognostic factors of esophageal squamous cell carcinoma.The risk-model formula is Riskscore=1.303×AL 033384.1-1.525×AC 108449.2,and the overall survival in the low risk group was longer than that of the high risk group(P<0.001).Principal component analysis showed the model-lncRNA could better distinguish the high and low risk groups,and it was the main influencing factor for the distribution in the two groups.Further GSEA enrichment analysis revealed gene functional annotation differences between the two groups,and the high risk group was mainly enriched in the process of keratinization and intracellular hypermetabolism.Besides,the 1-year,2-year and 3-year prediction performance of this risk-model was 0.750,0.768 and 0.796,respectively,which was better than those of other clinical characteris

关 键 词：食管鳞癌 长链非编码RNA 风险评分 预后 

分 类 号：R735.1[医药卫生—肿瘤]