A seven-DNA methylation signature as a novel prognostic biomarker in breast cancer.

Abstract

Breast cancer (BC) is a common malignant tumor and its incidence and mortality rates are ranked first among female cancers. So far, there has been no effective biomarkers for BC prognosis. The DNA methylation data of BC was downloaded from The Cancer Genome Atlas (TCGA), Gene Expression Omnibus, and Functional ANnoTation of The Mammalian Genome databases. The RNA-Seq data and clinical information of patients were downloaded from TCGA. R packages edgeR and minfi were used for differentially methylated genes (DMGs) screening. Then, the DMGs were collected for gene ontology and Kyoto Encyclopedia of Genes and Genomes analysis by the online tool database for annotation, visualization and integrated discovery (DAVID) and Reactome. Cox regression analysis was used to screen candidate differentially methylated sites (DMSs) for BC prognosis. Logrank test was used to explore the correlation between DMSs and survival time. Correlation analysis was used to investigate the correlation between DNA methylation and gene expression. We identified 276 DMGs which contained 1454 DMSs in those three datasets. Also, six DMGs that contained seven DMSs were identified by Cox regression analysis. Interestingly, their expression levels were negatively correlated with the DNA methylation level and not affected by age, subtypes, or tumor stages. We proposed that these seven differentially DNA methylation sites can be used as a novel prognostic biomarker for BC area under curve (AUC) = 0.74), which may facilitate research and benefit the clinical treatment of BC.