基于TCGA和GEO数据库结肠癌表观遗传预后模型构建与验证

Construction and validation of an epigenetic prognostic signature of colorectal cancer based on TCGA and GEO databases

目的使用TCGA和GEO数据库构建结肠癌(colorectal cancer,CRC)表观遗传的预后模型并验证。方法首先,公共数据库下载试验组数据(TCGA-COAD,n=446)、验证组数据(GSE39582,n=561)及表观遗传基因(EpiFactors,n=720);其次,单因素Cox回归筛选TCGA、GEO预后相关的基因,与EpiFactors数据库交集得到表观遗传相关影响预后的基因(epigenetic-related genes,ERGs)并对其进行富集分析;随后,多因素Cox回归构建预后模型,绘制列线图并验证。结果本研究共筛选得到6个ERGs,并构建预后模型,预测模型ROC的AUC为0.722,1年、3年和5年的校准曲线与实际结果相吻合,提示预测能力良好。结论本研究通过公共数据库筛选出6个ERGs(LBR、SMARCD3、ZNF532、MAPKAPK3、SFPQ、SIN3B)与结肠癌预后密切相关。

Objective To identify novel epigenetic-related biomarkers in colorectal cancer(CRC).Methods The training dataset was downloaded from The Cancer Genome Atlas(TCGA-COAD,n=446),the validation dataset was downloaded from the Gene Expression Omnibus(GSE39582,n=561),epigenetic-related genes were downloaded from the EpiFactors dataset(n=720).Overall survival-related genes were screened by Univariate Cox regression analysis in TCGA and GEO,epigenetic-related genes(ERGs)were identified by interacting those genes with EpiFactors,the enrichment analysis of ERGs also detected,then a prognostic signature was constructed by multivariate Cox regression analysis and validated.Results Six ERGs were screened,and the prognostic signature was constructed in our study,the AUC of the ROC was 0.722,and the calibration curve of 1-year,3-year,and 5-year were similar with the reality results,showed the signature had great prediction ability.Conclusion This study identified 6 ERGs(LBR,SMARCD3,ZNF532,MAPKAPK3,SFPQ,SIN3B),which are closely related to the prognosis of CRC.