Robin Andersson, et al., An atlas of active enhancers across human cell types and tissues, Nature 507, 2014
Here we use the FANTOM5 panel of samples, covering the majority of human tissues and cell types, to produce an atlas of active, in vivo-transcribed enhancers.
Enhancers were originally defined as remote elements that increase transcription independently of their orientation, position and distance to a promotor. They were only recently found to initiate RNA polymerase II transcription, producing so-called eRNAs.
Based upon the FANTOM5 CAGE expression atlas encompassing 432 primary cell, 135 tissue and 241 cell line samples from human, we identify 43,011 enhancer candidates and characterize their activity across the majority of human cell types and tissues.
RNA-seq data from matching primary cells and tissues showed that ~95% of RNAs originating from enhancers were unspoiled and typically short (median 346 nucleotides) – a striking difference to mRNAs. (19% unspoiled, median 1,256 nucleotides)
From the data we can draw several conclusions:
First, the majority of detected enhancers within any facet are not restricted to that facet.
Second, despite their apparent promiscuity, enhancers are more generally detected in a much smaller subset of samples than mRNA transcripts.
Third, the number of detected expressed enhancers and mRNA transcripts is correlated, but the number of detected expressed gene transcripts (>1 tag per million mapped reads(TPM)) is 19-34 fold larger than the number of detected enhancers with the cut-offs used. Noteworthy exceptions include blood and immune cells, testis, thymus and spleen, which have high enhancers/gene ratios. Conversely, smooth and skeletal muscle and skin, bone and epithelia-related cells have low ratios.
Hierarchical clustering of enhancers by facet expression revealed a small subset of enhancers (200 or 247, defined by primary cell or tissue facets, respectively) expressed in the large majority of facets.
U-enhancers comprise a small but distinct subset of enhancers, which probably has specific regulatory functions used by virtually every human cell.
Disease-associated SNPs were over-represented in regulatory regions to a greater extent than in exons.