Supplementary Figure 18 : Comparisons with published silencers

Supplementary Figure 18 | Comparisons with published silencers

a, Heatmap depicting shared cCREs across ENCODE silencer categories (REST+ enhancer/silencer, REST+ silencer, STARR silencer (stringent), and STARR silencer (robust) cCREs) and cCREs that overlap previously published silencer annotations from Cai et al., Huan et al., Jayavelu et al., Pang et al. b, Barplots depicting the percentage of silencer cCREs tested by Jayavelu et al. that were deemed to have silencer activity in that study. Results for all silencer cCREs are in black, REST+ enhancer/silencer cCREs in yellow, REST+ silencer cCREs in purple, stringent STARR silencer cCREs in dark blue, and robust STARR silencer cCREs in light blue. c, Percentage of ENCODE silencer cCREs tested by Jayavelu et al. stratified as described in b. d, Percentage of ENCODE silencer cCREs with high DNase signal in K562 cells, stratified as described in b. e, Heatmap showing the fraction of each silencer cCRE group (as defined in a), K562 promoter, K562 distal enhancer, and cCEs with low chromatin accessibility in K562 that overlap K562 ChromHMM states. State details are described in Supplementary Table 14c. f, Heatscatter plot showing the distribution of cCREs that overlap Huan et al. silencers based on their alignment across Zoonomia mammalian genomes. N1 denotes the number of species in which ≥90% of a cCRE aligns; N2 denotes the number of species in which ≤10% of a cCRE aligns. Rectangles indicate predefined groups: Group 1 (G1), highly conserved (N1 ≥ 120, N2 ≤ 25); Group 2 (G2), actively evolving (20 ≤ N1 ≤ 50, N2 ≤ 120); and Group 3 (G3), primate-specific (N1 ≤ 50, N2 ≥ 180). g, Heatscatter plot as in f, but for cCREs that overlap Cai et al. h, Heatscatter plot as in f, but for cCREs that overlap Pang et al. i, Heatscatter plot as in f, but for cCREs that overlap Jayavelu et al.