Fig. 2
From: Clustered CTCF binding is an evolutionary mechanism to maintain topologically associating domains
CTCF binding sites at TAD boundaries are subjected to stronger evolutionary constraints. a CTCF-bound sites at TAD boundaries contain motifs with a higher binding affinity for CTCF than non-TAD boundary-associated sites (Mann-Whitney U test: p value < 2.2e−10). b Although the binding affinity of CTCF sites is generally proportional to the conservation level of the site (how many species it is shared by), CTCF sites at TAD boundaries have stronger binding affinity than non-TAD boundary-associated sites, independent of their conservation level (Mann-Whitney U tests between TAD boundary-associated and non-TAD boundary-associated sites: p1-way = 0.001, p2-way = 0.06, p3-way = 6.1e−07, p4-way = 5.2e−13, p5-way = 3.9e−11). c TAD boundary-associated CTCF peaks display higher ChIP enrichment scores, as calculated by MACS, than non-TAD boundary-associated peaks (Mann-Whitney U test: p value < 2.2e−10). d TAD boundary-associated CTCF peaks, at every conservation level, display stronger ChIP enrichment than non-TAD boundary-associated peaks (Mann-Whitney U tests: p1-way < 2.2e−16, p2-way = 0.002316, p3-way < 2.2e−16, p4-way < 2.2e−16, p5-way = 2.047e−12). e The most information-rich bases of the primary CTCF M1 motif at TAD boundaries display higher rejected substitution (RS) scores compared to non-TAD boundary-associated motifs. The bottom panel shows the position weight matrix of the CTCF M1 motif from Schmidt et al. [34] f The observation in e is independent of the conservation level of the CTCF sites, as shown for subsets of sites at each conservation level