Fig. 3

Co-localization of TAD boundaries and synteny breaks. a Representative chromatin interaction map involving a genomic region (750 kb) in KI_01 of G. kirkii, in which the component TADs are outlined by diagonal rectangles. b Fractions of TAD boundaries overlapping with synteny breaks identified in comparisons of G. kirkii vs. G. arboreum and G. kirkii vs. G. raimondii are statistically higher than those in multiple randomization controls, which involve groups of shuffled TAD boundaries (identified synteny breaks are maintained), shuffled synteny breaks (identified TAD boundaries are maintained), and both TAD boundaries and synteny breaks shuffled simultaneously. The maximum p value of Fisher’s two-tailed test in comparison to the respective control is still less than 0.001. c Fractions of TAD boundaries co-localizing with synteny breaks within TAD categories of different sizes in G. kirkii, G. arboreum, and G. raimondii, based on the regression between fraction and TAD size. d DNA methylation levels (in CG, CHG, and CHH contexts) in boundaries of TAD groups (large and small TADs) in G. kirkii, G. arboreum, and G. raimondii, respectively. e Abundance of active histone modification (H3K4me3) and transcribed RNA-seq reads in boundaries of TAD groups (large and small TADs) in G. kirkii, G. arboreum, and G. raimondii, respectively. Specific numbers of TADs in respective large and small TAD group are parenthesized at the bottom of each panel. Statistical significance was calculated using Wilcoxon’s rank sum test