Integrated DNA methylation and gene expression analysis identifies CpG biomarkers in breast cancer

DNA methylation alterations are frequently associated with transcriptional dysregulation and tumor-associated molecular changes in breast cancer; however, CpG-specific methylation patterns linked to aberrant gene expression remain incompletely characterized. In this study, we performed an integrative analysis of genome-wide DNA methylation and transcriptomic data from the TCGA breast cancer cohort to identify methylation-associated transcriptional signatures and candidate biomarkers. DNA methylation profiles generated using the Illumina HumanMethylation450K platform were integrated with RNA-sequencing data using promoter-focused correlation analysis (798 tumour and 97 normal samples). Differential methylation and differential expression analyses identified coordinated epigenetic and transcriptional alterations between tumor and normal tissues. Machine-learning trained on selected CpG signatures achieved strong classification performance, with Random Forest performing best, in both the internal TCGA cohort (798 tumour and 97 normal samples) and an independent external validation cohort from GEO (295 tumor and 100 normal samples). Functional enrichment analysis revealed significant enrichment of immune-related and metabolic pathways associated with breast cancer progression. Experimental validation by RT-qPCR further confirmed significant dysregulation of KLK2 expression in breast tumor tissues. Collectively, these findings identify CpG-associated methylation signatures linked to transcriptional dysregulation in breast cancer and highlight their potential utility as candidate epigenetic biomarkers.