Extended Data Figure 2: Enhancers have a complex interaction network with multiple active enhancers and promoters.
From: Enhancer loops appear stable during development and are associated with paused polymerase
a, Histogram showing the frequency of viewpoint interactions, averaged over all conditions per viewpoint (n = 107). b, Frequency of 4C unique interacting regions that overlap enhancers or promoters when using distal enhancers (left) or promoter-proximal elements (right) as viewpoints (defined in Extended Data Fig. 1a). Whole-embryo RNA-seq9 was used to define active and inactive promoters. Enrichments over a background set of interactions (Fisher’s Exact Test) are indicated. The background set (or expectation regions) was based on random sampling of all DpnII fragments throughout the genome that have matched mappability, G+C content and size to the observed interactions. c–e, Mesoderm-specific chromatin signatures6 at whole embryo 6–8 h 4C interacting regions. c, H3K27ac, H3K4me3, H3K79me3 and Pol II signals6 are enriched at promoters interacting with promoter-proximal viewpoints. Note, we only consider interactions between promoter-proximal viewpoints and distal promoters (>1 kb away; Extended Data Fig. 1a). d, H3K27ac, H3K4me1, H3K79me3 and Pol II signals6 are enriched at enhancers interacting with promoter-proximal viewpoints. e, H3K27ac, H3K4me3, H3K79me3 and Pol II signals6 are enriched at promoters interacting with distal enhancer viewpoints. Shading indicates 95% confidence intervals estimated by non-parametric bootstrapping, grey lines indicate background signal.
