A) Global biochemical profiles support a coherent model for cEnhs. Distal and proximal candidate enhancer elements shown. Non H3K27Ac modified histones are represented by the light blue core histones. P300 (dark grey) is recruited either directly by a TF, or through cofactors to acetylate adjacent histones (dark blue core), which are displaced, creating in most cases DNAse hypersensitive (purple DNA) regions where the TF is occupying the DNA.
B) The transcription factor (TF) occupancy, DNAse hypersensitivity (DHS) and H3K27ac global summed signatures are strikingly similar at both distal and proximal cEnhs. DNAse-seq and ChIP-seq experiments against H3K27ac and the TF myogenin in differentiated (myocyte) C2C12 cells[red]; in Gata1 restored G1E-ER4 erythroid cells [blue] and in GR activated (dex induced) A549 cells [black] were analyzed. The 1kb radius around the ChIP-Seq peak of statistically significant sites (IDR=0.05) were mapped and genome wide bp by bp density over the window is shown for the TF; DNAse and H3K27Ac measurements for TSS proximal (within 2kb) and distal candidate elements in the genome. Both H3K27Ac modified and unmodified histones largely protect the DNA from DNAse. 
C) Global TF measurements detect significant sites (IDR>0.05) that vary in signal intensity over 2+logs. DNAse-seq and ChIP-seq experiments against the TF MyoD, H3K27ac, p300 in C2C12 myocyte cells were analyzed. The 1kb radius region around the ChIP-Seq peak of all 32,278 statistically significant sites (IDR=0.05) were mapped and the top, mid, lowest 500 signal sites are shown. The signal for DNAse, H3K27Ac and p300 measurements for those regions are shown along with instances of the DNA motif myogenin has been shown to bind (muscle E-box RRCAGSTG).
D) The vast majority of distal cEnhs use cell type differential locations in the genome, while the majority of TSS-proximal cEnhs are shared among cell types. Statistically significant DHS (IDR=0.05) from HepG2 (light grey) and K562 (dark grey) were analyzed and split into distal and TSS proximal (XXXbp radius). The DHS sites were then overlapped and compared. In total 53704 distal DHS sites were called of which only 5205 (9.6%) were shared (black). 23492 TSS Proximal DHS sites were found, of which 11,181 (47.6%) were shared between the two cell types. 
E) Outline of experimental design, candidate Enhancer (cENH) selection and enhancer assay. cEnhs were selected based on TF ChIP-Seq occupancy measurements in Muscle (C2C12 – Myogenin/Myod), Erythroid (G1E-ER4 – GATA1/Tal1) and in immortalized K562 cells (multiple TFs). Machine learning based cEnhs were selected in K562 cells based on single tissue chromatin segmentation (chromHMM, segway) prediction while liver (HepG2) specific cEnhs were selected from a SOM that integrated chromatin states across multiple tissues. GR ChIP elements were cloned into a Starr-Seq libeary and assayed for activity. All candidate enhancers were tested in the corresponding cell lines, except the Erythroid mouse elements, which were tested in human K562 cells for activity; then scored for DNAse, TFs and histone marks.