The majority of all TF binding events occur at <10% of all TF sites and can be largely predicted from as few as 30 TFs. (A) Merged TF binding sites in HepG2 cells ranked on decreasing TF enrichment (number of the overlapping TF peaks; red line) and the corresponding cumulative frequency curve (blue line). (B) From zero to 80% cumulative frequency increased at 10% intervals, the average H3K27ac signal intensity is plotted for the merged TF binding sites at each interval. Green lines mark the points at 10%, 20%, 30%, 40%, 50%, 60%, 70%, and 80% of cumulative levels. (C) H3K27ac peak sites ranked on decreasing TF enrichment (red line) and the corresponding accumulative frequency curve (blue line). Peak index at 60% and 80% of accumulative levels is marked. (D) From two randomly selected nonoverlapping pools of n TFs (n varies from 30–90), we identified the top 15,000 most TF-occupied targets (MOTs), as well as the top 15,000 MOTs obtained from all 195 TFs in HepG2 cells. The MOTs are marked with arrows. (E) The overlaps of the top 15,000 MOTs between two randomly selected nonoverlapping pools of n TFs (n = 30, 40, 50, 60, 70, 80, 90). Each box in the plot represents results from 100 times of random TF selection. (F) The overlaps of these MOTs and the top 15,000 MOTs obtained from all 193 TFs.
