Improving representations of genomic sequence motifs in convolutional networks with exponential activations

Genome-scale studies have revealed extensive, cell type-specific colocalization of transcription factors, but the mechanisms underlying this phenomenon remain poorly understood. Here, we demonstrate in macrophages and B cells that collaborative interactions of the common factor PU.1 with small sets of macrophage- or B cell lineage-determining transcription factors establish cell-specific binding sites that are associated with the majority of promoter-distal H3K4me1-marked genomic regions. PU.1 binding initiates nucleosome remodeling, followed by H3K4 monomethylation at large numbers of genomic regions associated with both broadly and specifically expressed genes. These locations serve as beacons for additional factors, exemplified by liver X receptors, which drive both cell-specific gene expression and signal-dependent responses. Together with analyses of transcription factor binding and H3K4me1 patterns in other cell types, these studies suggest that simple combinations of lineage-determining transcription factors can specify the genomic sites ultimately responsible for both cell identity and cell type-specific responses to diverse signaling inputs. Copyright 2010 Elsevier Inc. All rights reserved.

Summary The human genome encodes the blueprint of life, but the function of the vast majority of its nearly three billion bases is unknown. The Encyclopedia of DNA Elements (ENCODE) project has systematically mapped regions of transcription, transcription factor association, chromatin structure, and histone modification. These data enabled us to assign biochemical functions for 80% of the genome, in particular outside of the well-studied protein-coding regions. Many discovered candidate regulatory elements are physically associated with one another and with expressed genes, providing new insights into the mechanisms of gene regulation. The newly identified elements also show a statistical correspondence to sequence variants linked to human disease, and can thereby guide interpretation of this variation. Overall the project provides new insights into the organization and regulation of our genes and genome, and an expansive resource of functional annotations for biomedical research.

The reference human genome sequence set the stage for studies of genetic variation and its association with human disease, but a similar reference has lacked for epigenomic studies. To address this need, the NIH Roadmap Epigenomics Consortium generated the largest collection to-date of human epigenomes for primary cells and tissues. Here, we describe the integrative analysis of 111 reference human epigenomes generated as part of the program, profiled for histone modification patterns, DNA accessibility, DNA methylation, and RNA expression. We establish global maps of regulatory elements, define regulatory modules of coordinated activity, and their likely activators and repressors. We show that disease and trait-associated genetic variants are enriched in tissue-specific epigenomic marks, revealing biologically-relevant cell types for diverse human traits, and providing a resource for interpreting the molecular basis of human disease. Our results demonstrate the central role of epigenomic information for understanding gene regulation, cellular differentiation, and human disease.