Fig. 3

High-order chromatin organization and its transcriptional implications. a Overlapping of different types of histone modification and RNAPII determined anchor peaks. The number of anchor peaks for a given histone modification and RNAPII are shown on the left. The percentages of overlap between two indicated factor determined anchor peaks is represented by the area of the circle. b, c Expression levels among genes with different binding (basal) and interaction (anchor) patterns. ***p < 2.2e-16 from Wilcoxon test. d Bar chart of genomic coverage by different chromatin interacting domains (CIDs). CIDs cover 71% of the Arabidopsis genome, whereas a minor portion is represented by gap regions with no covering by chromatin interactions (35 Mb, 29%). AID (active interacting domains), 47 Mb, 39%; RID (repressive interacting domains), 15 Mb, 12%; HID (heterochromatin interacting domains), 17 Mb, 14%; and MID (mix interacting domains), 6 Mb, 5%. e Browser view of different CIDs showing interval arrangement. The data tracks show CIDs, chromatin loops, profiling of representative histone modifications and RNAPII occupancy, and gene transcription. Within the CID-associated genomic landscape, different CID module categories represent distinct transcriptional activities. f Percentage of RNA-seq reads mapped to the indicated CID regions. g Boxplot for transcription levels of anchor genes in four CID categories. Genes from gap regions served as control. p < 2.2e-16 from Kruskal–Wallis test. h Gene density of the AID and HID in Arabidopsis and rice. Genome, gene density of genome. ns, no significant difference. ***p < 2.2e-16 from Wilcoxon test