基于染色质调控因子的乳腺癌预后标志物筛选和风险模型构建

Biomarker Screening and Construction of a Prognostic Risk Model Based on Chromatin Regulators in Breast Cancer

染色质调控因子(chromatin regulators, CRs)是肿瘤发展过程中必不可少的上游调控因子。本研究旨在筛选在乳腺癌中与CRs相关的风险标志物并构建相应的预后风险模型,探讨CRs在乳腺癌预后中的作用和价值。从癌症基因组图谱(The Cancer Genome Atlas, TCGA)数据库中下载乳腺癌及癌旁组织样本的转录组数据、临床数据,从FACER数据库中下载染色质调控作用相关的基因集,取交集后进行差异表达分析得到差异CRs基因集;将有效临床样本分为训练集和验证集,在训练集中利用单因素COX回归、 LASSO回归、多因素COX回归进行特征筛选,并构建了预后风险模型,分别进行Kaplan-Meier生存分析、 ROC分析以评价模型,同时在验证集中进行验证,利用单因素及多因素COX回归分析判断其是否为独立预后因素;对风险模型特征进行GO与KEGG功能富集、免疫浸润、药物敏感性等分析。最终筛选到9个含显著性风险的特征因子,分别是ASCL2、EXOSC4、IDH2、IKZF3、MECOM、PRDM12、PRMT8、TDRKH、TFF1,并由这些因子构建预后风险模型,通过生存和预后相关性等分析得到该模型特征因子与临床因素具有显著相关性,其他综合分析得知风险模型因子PRDM12、MECOM的风险特征最为显著。由这9个CRs基因组成的风险模型具有良好的预后预测能力,其特征因子可作为乳腺癌参考风险生物标志物。

Chromatin regulators(CRs)are essential upstream regulators in the process of tumor development.The aims of this study were to screen risk markers associated with CRs in breast cancer and construct corresponding prognostic risk models to investigate the role and value of CRs in the prognosis of breast cancer.The transcriptomic data and clinical data of breast cancer and paraneoplastic tissue samples were downloaded from The Cancer Genome Atlas(TCGA)database,and the gene sets related to chromatin regulatory effects were downloaded from FACER database,and differential expression analysis was performed after taking the intersection to obtain differential CRs gene sets.Valid clinical samples were then divided into two datasets:training set and test set.In the training set,sin--gle-factor COX regression,LASSO regression and multi-factor COX regression were used for the feature screening,and a prognostic risk model was constructed.Kaplan-Meier survival analysis and ROC analysis were performed to evaluate the model,respectively.Valida-tion was also performed in the test set.Then,used single-factor and multi-factor COX regression analysis to determine whether it was an independent prognostic factor.CO and KEGG enrichment analyses were performed on the risk model features,functional enrich-ment,immune infiltration,and drug sensitivity were analyzed.Finally,nine significant risk characteristic factors were picked out,which were ASCL2,EXOSC4,IDH2,IKZF3,MECOM,PRDM/2,PRMT8.TDRKH,TFFI,and these factors constitute the prognostic risk model.Through the survival and prognostic correlation analysis,it was found that the model was significantly correlated with the clinical factors.Furthermore,it was found that the risk model factors PRDMI2 and MECOM were the most significant risk characteristics.The risk model composed of these 9 CRs genes has good prognostic predictive ability,and its characteristic factors can be used as reference risk biomarkers for breast cancer.