基于TCGA数据库探究ESCRT相关基因对膀胱癌的潜在预后价值  

An Exploratory Study of the Potential Prognostic Value of ESCRT-related Genes for Bladder Cancer Based on TCGA Database

作　　者：陈晓睿 王晓晨 张宗亮[3] 原江水[2] 宋卫青[2] CHEN Xiaorui;WANG Xiaochen;ZHANG Zongliang;YUAN Jiangshui;SONG Weiqing(Qingdao University,Qingdao 266011,China;Department of Laboratory Science,Qingdao Municipal Hospital,Qingdao 266011,China;Department of Urology,Affiliatedee Hospital of Qingdao University,Qingdao 266011,China)

机构地区：[1]青岛大学,山东青岛266011 [2]青岛市市立医院检验科,山东青岛266011 [3]青岛大学附属医院泌尿外科,山东青岛266011

出　　处：《标记免疫分析与临床》2023年第9期1531-1538,1620,共9页Labeled Immunoassays and Clinical Medicine

基　　金：国家自然科学基金面上项目(编号:31971191)。

摘　　要：目的 通过癌症基因图谱(TCGA)分析内吞体分选转运复合体(ESCRT)相关基因对膀胱癌预后的潜在预测价值,构建和评估膀胱癌预后模型。方法 通过访问TCGA数据库获取膀胱癌的临床病例资料和转录组数据。筛选与ESCRT相关基因在膀胱肿瘤组织与正常组织中的差异表达基因进行相关功能学富集分析,构建蛋白相互作用网络,采用Lasso和Cox回归方法,筛选与患者总生存期密切相关的基因,基于这些基因进一步构建出患者预后风险评分模型,并评估其预测能力。结果 从TCGA数据库下载包含405个膀胱癌组织和19个正常组织的RNA-seq信息,通过差异分析筛选出6个重合基因即膀胱癌差异表达的ESCRT家族基因,经过单因素Cox回归分析,发现存在3个基因对膀胱癌患者的预后有显著的影响。通过Lasso和Cox回归筛选分析最终得到2个(MVB12B、CHMP4C)与膀胱癌预后相关的关键基因并以此构建预后模型,预测训练集和验证集的1年、3年和5年ROC曲线下面积分别为0.768、0.694、0.732和0.651、0.720、0.776。结论 成功构建了基于2个关键DEGs表达的膀胱癌预后风险预测模型,该模型可为预测膀胱癌患者的预后提供候选生物标志物,并为临床中评估膀胱癌患者的预后提供有价值的参考。Objective The bioinformatics method was used to mine the cancer gene map(TCGA)to explore the potential predictive value of transport-essential endosome sorting complex(ESCRT)-related genes for the prognosis of bladder cancer(BLCA)and to construct and evaluate the prognosis model of bladder cancer.Methods BLCA clinical data from the TCGA database and transcriptome data were used for the study.We screened ESCRT-related genes in bladder tumor tissue and tissue adjacent to carcinoma of differentially expressed genes(DEGs)for enrichment related function analysis.We constructed protein interaction networks(Protein-Protein Interaction Networks,PPI),and then used Lasso and Cox regression to identify patients’overall survival of genes,and their prognosis value with a risk score model,and further evaluated its ability for prediction.Results The RNA-Seq information for 405 BLCAs and 19 para-cancerous tissues was downloaded from the TCGA database.Six coincidence genes,namely ESCRT family genes with differential expression in bladder cancer,were screened out by differential analysis.Univariate Cox regression analysis showed that three genes had a significant influence on the prognosis of bladder cancer.Lasso and Cox regression screening analysis yielded two key genes(MVB12B,CHMP4C)associated with bladder cancer prognosis and we used them to construct a prognostic model.The areas under the ROC curves for one,three,and five years in the training and validation sets were 0.768,0.694,0.732,and 0.651,0.720 and 0.776,respectively.Conclusion This study successfully constructed a prognostic risk prediction model for bladder cancer based on two key genes expressions.This model can effectively predict the prognosis of patients with bladder cancer and provide candidate biomarkers for prognosis,which has reference value for clinical prediction of the prognosis of patients with bladder cancer and,to a certain extent,provides a novel diagnosis and treatment strategy.

关 键 词：膀胱癌 ESCRT CHMP4C MVB12B 预后 

分 类 号：R737.14[医药卫生—肿瘤]