Identification of the genomic regions that regulate transcription remains an important open problem. We have recently shown that global run-on and sequencing (GRO-seq) with enrichment for 5'-capped RNAs reveals patterns of divergent transcription that accurately mark active transcriptional regulatory elements (TREs), including enhancers and promoters. Here, we demonstrate that active TREs can be identified with comparable accuracy by applying sensitive machine-learning methods to standard GRO-seq and PRO-seq data, allowing TREs to be assayed together with transcription levels, elongation rates, and other transcriptional features, in a single experiment. Our method, called discriminative Regulatory Element detection from GRO-seq (dREG), summarizes GRO-seq read counts at multiple scales and uses support vector regression to predict active TREs. The predicted TREs are strongly enriched for marks associated with functional elements, including H3K27ac, transcription factor binding sites, eQTLs, and GWAS-associated SNPs. Using dREG, we survey TREs in eight cell types and provide new insights into global patterns of TRE assembly and function.

28Transcriptional regulatory changes have been shown to contribute to phenotypic differences 29 between species, but many questions remain about how gene expression evolves. Here we report 30 the first comparative study of nascent transcription in primates. We used PRO-seq to map actively 31 transcribing RNA polymerases in resting and activated CD4+ T-cells in multiple human, 32 chimpanzee, and rhesus macaque individuals, with rodents as outgroups. This approach allowed us 33 to measure transcription separately from post-transcriptional processes. We observed general 34 conservation in coding and non-coding transcription, punctuated by numerous differences between 35 species, particularly at distal enhancers and non-coding RNAs. We found evidence that 36 transcription factor binding sites are a primary determinant of transcriptional differences between 37 species, that stabilizing selection maintains gene expression levels despite frequent changes at 38 distal enhancers, and that adaptive substitutions have driven lineage-specific transcription. Finally, 39we found strong correlations between evolutionary rates and long-range chromatin interactions. 40These observations clarify the role of primary transcription in regulatory evolution. 41 . CC-BY 4.0 International license peer-reviewed) is the author/funder. It is made available under a The copyright holder for this preprint (which was not . http://dx.doi.org/10.1101/083212 doi: bioRxiv preprint first posted online Oct. 25, 2016; 1 Following decades of speculation that changes in the regulation of genes could be a potent 42 force in the evolution of form and function [1][2][3] , investigators have now empirically demonstrated the 43 evolutionary importance of gene regulation across the tree of life [4][5][6][7][8][9][10][11][12] . Changes in gene expression 44 are primarily driven by mutations to non-coding DNA sequences, particularly those that bind 45 sequence-specific transcription factors 13 . Accordingly, adaptive nucleotide substitutions at 46 transcription factor binding sites (TFBSs) 9,10,[14][15][16] and gains and losses of TFBSs 17-25 both appear to 47 make major contributions to the evolution of gene expression. These events are believed to modify 48 a variety of rate-limiting steps early in transcriptional activation 26 . In addition, transcriptional 49 activity is generally correlated with various epigenomic and structural features, including post-50 translational modifications to core histones, the locations of architectural proteins such as CTCF, 51 and the organization of topological associated domains. Like TFBSs, these features display general 52 conservation across species, yet do exhibit some variation, with differences between species 53 roughly proportional to evolutionary distance 27 . Moreover, differences between species in these 54 features correlate with differences in gene expression 8,24,[28][29][30] . 55Nevertheless, many open questions remain about the roles of TFBSs, chromatin 56 organization, and posttranscription...

No abstract