Fig. 5

Genome-wide analysis of ARID1A-dependent CHD4 binding. A MA plot of shARID1A vs. control differential CHD4 ChIP-seq (n = 2), across 44,567 tested genomic regions of CHD4 binding. Regions are colored based on shARID1A differential CHD4 binding significance. Inset pie chart depicts distribution of significantly increasing and decreasing CHD4 regions (csaw/edgeR FDR < 0.05) compared to stable CHD4 (FDR > 0.05). FDR < 0.05 was used as the significance threshold for all downstream analyses. B Global analysis of ARID1A-dependent CHD4 binding based on presence of normal ARID1A binding. Box plot quantification of shARID1A log₂FC CHD4 binding, segregated by ARID1A binding status. Statistic is two-tailed, unpaired Wilcoxon’s test. C Chromatin state enrichment among shARID1A increasing and decreasing CHD4 binding regions, calculated as observed/expected genomic fold-enrichment per genomic bp. Statistic is hypergeometric enrichment. D Enrichment of ARID1A binding detection at regions with decreasing CHD4 binding following ARID1A loss compared to all tested CHD4 binding sites. Statistics are hypergeometric enrichment test and pairwise two-tailed Fisher’s exact test. E Enrichment of ARID1A-dependent H3.3 maintenance (shARID1A decreasing H3.3 abundance) at regions with decreasing CHD4 binding following ARID1A loss compared to all tested CHD4 binding sites. Statistics are hypergeometric enrichment test and pairwise two-tailed Fisher’s exact test. F Associations between presence of H3.3 (wild-type peak) and ARID1A-bound, ARID1A-dependent CHD4 binding. Statistic is hypergeometric enrichment. G Associations between ZMYND8 co-binding (wild-type peak) and ARID1A-bound, ARID1A-dependent CHD4 binding. Statistic is hypergeometric enrichment. *** p < 0.001