Table 1.

Results of MethReg stage-wise analysis of TCGA COADREAD dataset

Analysis	Unique triplets	Unique CpGs	Unique TFs	Unique targets
Promoter analysis
Screening	165	49	56	42
Confirmation	31	29	18	25
Distal analysis
Screening	2358	315	155	570
Confirmation	52	43	44	44
Regulon-based analysis
Screening	78	23	19	19
Confirmation	47	18	11	6

After removing effects of copy number alteration and tumor purity estimate in gene expression data, the robust linear model target gene expression residual ∼ TF activity +DNAm.group +DNAm.group× TF activity was used to analyze CpG–TF–target gene triplets. In each analysis, TFs with binding sites within 250 bp of the CpGs were analyzed, and TF activity scores were estimated using the VIPER algorithm. Shown are significant triplets at 5% overall FDR level and the unique CpGs, TFs and targets in these triplets. The CpGs in the promoter regions (±2 kb around the TSS) were linked to genes that had promoters overlapping with them (promoter analysis), while CpGs in the distal regions (>2 kb from TSS) were linked to five genes upstream and five genes downstream within 1 million bp (distal analysis). In regulon-based analysis, CpGs were linked to TF–target gene pairs in the pan-cancer regulons described in (31). Abbreviations: DNAm, DNA methylation; TSS, transcriptional start site; TF, transcription factor