Fig. 7

Analysis of cell type-specific enhancers defined by enhancer transcription. a Fraction of cell type-specific (n = 1,052) and non-cell type-specific (n = 837) enhancer transcript pairs in the universe of all enhancer transcript pairs called from GRO-seq data using groHMM across four different cell lines (MCF-7, LNCaP, IMR90, AC16). b Distribution of enhancer transcript pairs from (A) in one or more cell types. Transcript pairs in one cell type that overlapped a transcript pair in one or more other cell types by at least 20% of their length were counted and summed. c Heatmaps showing the relative expression of 1,052 cell type-specific enhancer transcript pairs (top) and 837 non-cell type-specific enhancer transcript pairs (bottom) after hierarchical clustering analysis. The hierarchical clustering analysis was performed on both the rows and columns using GRO-seq reads on both strands for each enhancer (Ward’s method; [57]). d Heatmaps showing normalized GRO-seq read counts for 1,052 cell type-specific enhancer transcript pairs (top) and 837 non-cell type-specific enhancer transcript pairs (bottom). The order of the enhancers from top to bottom is the same as in (C). Two hundred and fifty bp windows within 10 kb regions from the center of the enhancers are shown. e Metagene representations showing the average GRO-seq read distributions ± 4 kb around the center of the enhancer transcript pair overlap for cell type-specific enhancers in their cognate cell type (e.g., MCF-7 cell-specific enhancers in MCF-7 cells) (top) or cell type-specific enhancers in other cell types (e.g., LNCaP-, IMR90-, and AC16-specific enhancers in MCF-7 cells) (bottom)