Fig. 1

The complete workflow of the study. (I): Analysis initially used three DNA methylation datasets (Absher, Kirby and TCGA) and the Meta expression dataset of gene expression. Based on promoter methylation status and gene expression, four groups of gene regulation patterns were established, and genes were grouped accordingly. (II): The 450 K TCGA DNA methylation dataset was used to associate genes with more methylation probes to filter into UPUP-only and UPDOWN-only gene groups by removing genes with ambiguous DNA methylation statuses. (III): Correlation analysis was performed using TCGA combined dataset, where DNA methylation and gene expression values derived from the same samples were individually compared. (IV): UPUP-only and UPDOWN-only genes were studied genome-wide by analysing the distances between probes and TSSs of the associated genes, genomic locations of probes, and performing a GSEA. In addition, several individual genes were investigated by visualizing methylation probes