Extended Data Figure 7: Identification of enhancer–promoter interactions.
From: Chromatin architecture reorganization during stem cell differentiation
a, Shown is the Hi-C interaction frequency between the IDO1 gene promoter regions and ± 1 Mb surrounding regions. Each entry in the heat map of Hi-C inter. freq. indicates Hi-C interaction frequency between the promoter and the surrounding regions. Each row indicates the Hi-C interaction frequencies for a given window size. The heat map of random permutation was generated by randomizing each row in Hi-C interaction frequency. The sum (Hi-C) and sum (random) indicate summation of Hi-C interaction frequencies for each 5-kb window. Predicted enhancers, H3K27ac, RNA-seq, and RefSeq gene information are shown together. b, Box plots showing the reproducibility of predicted enhancer–promoter interactions between two biological replicates for each chromosome with different P value cutoffs. If the P value is less than 0.001, the reproducibility between replicates is over 80%. c, Distribution of 5C signals between interacting pairs (interaction) and non-interacting pairs (others) defined by Hi-C interaction frequency score with different P value cutoffs. Interaction pairs defined by Hi-C interactions are also strongly enriched by 5C signals at both P value cutoffs (n = 11,461 for 1 × 10−3 and n = 1,841 for 1 × 10−4). d, Relationship between Hi-C interaction frequency scores and dsQTL target-gene pairs according to distance between gene and its target DHS regions. Target-gene relationships tend to show higher Hi-C interaction frequency scores compared to off-target-gene relationships.
