Figure 2.

Integrating Islet CAGE TCs with other epigenomic information reveals characteristics of transcription initiation. A: Aggregate CAGE profiles over ATAC-seq peak summits. B: Aggregate ATAC-seq profile over TC midpoints. C: Enrichment of TF footprint motifs to overlap 500 bp upstream region (y-axis) vs. 500 bp downstream region (x-axis) of islet TCs. Colors denote whether a TF footprint motif was significantly enriched (5% FDR correction, Benjamini-Yekutieli method) to overlap only upstream regions, only downstream regions, both, or none. D: Chromatin state annotations across 98 Roadmap Epigenomics cell types (18-state “extended model”) (37) for TC segments that occur in islet promoter chromatin states (11-state model) and overlap ATAC-seq peaks. These segments were segregated into those occurring 5 kb proximal (left) (N = 7,064 TC segments) and distal (right) (N = 443 TC segments) to known protein-coding gene TSS (GENCODE V19). E: Chromatin state annotations across 98 Roadmap Epigenomics cell types (18-state extended model) for TC segments that occur in islet enhancer chromatin states (11-state model) and overlap ATAC-seq peaks, segregated into those occurring 5 kb proximal (left) (N = 254 TC segments) and distal (right) (N = 289 TC segments) to known protein-coding gene TSS. Note that the heat map widths in D and E are scaled to aid in interpretability. F: Enrichment of TF footprint motifs to overlap TCs occurring in accessible enhancer chromatin states (y-axis) vs. TCs occurring in accessible promoter chromatin states (x-axis). Colors denote whether a TF footprint-motif was significantly enriched (5% FDR correction, Benjamini-Yekutieli method) to overlap only TCs in accessible enhancer regions, only TCs in accessible promoter regions, both, or none. G: Aggregate CAGE profiles centered and oriented relative to RFX5_known8 footprint motifs occurring in 5 kb TSS distal regions. H: Aggregate CAGE profiles centered and oriented relative to ELK4_1 footprint motifs.