Fig. 1

Identification and analysis of emTFs. A Box plot depicting the fraction of CpGs close to TFBSs for each TF (each point corresponds to a TF in a given cohort (columns)) with DNA methylation level correlated with the TF expression. The horizontal blue line represents the 95th percentile of the distribution of all fractions (see Additional file 3: Fig. S1C for the distribution). B Upset plot representing the emTF predictions across the cancer types. Each row represents a cancer type with points providing information about the intersection of the TFs predicted in the different cancer types. The bars at the top indicate the number of intersecting TFs (annotated above each bar) in each combination of cancer types (indicated by the points). The set size below the horizontal bars depicts the number of TFs predicted in each cohort. C For each emTF (columns), the plot provides the percentage of positively (purple dots) and negatively (green dots) correlated CpGs (emCpGs; y-axis) close to their TFBSs predicted in each cancer type (one per facet). See Additional file 4: Table S3 for the detailed numbers. D Venn diagram of the intersection between the predicted emTFs (n = 13), pioneer TFs from the literature (known PTFs; n = 34), and flanking accessibility-associated TFs from [43] (n = 29)