SRA

Knowledge of the locations and activities of cis-regulatory elements (CREs) is needed to decipher basic mechanisms of gene regulation and to better understand the impact of genetic variants on complex traits. Previous studies identified candidate CREs (cCREs) using multiple epigenetic features in cell types from one species at a time, but such an approach can make it difficult to compare results across species. In contrast, we conducted a cross-species study defining epigenetic states and identifying cCREs in blood cell types to generate maps of the regulatory landscapes and cCREs that are comparable across species. This study used integrative modeling of eight epigenetic features jointly in both human and mouse in our Validated Systematic Integration (VISION) Project. The accuracy of the cCRE predictions was supported by orthogonal function-related data. The contribution of each epigenetic state in cCREs to gene regulation was estimated from a multivariate regression against gene expression across cell types, and these values were used to estimate an epigenetic state Regulatory Potential (esRP) score for each cCRE in each cell type. These esRP scores have broad utility for visualizing and categorizing the dynamic changes in cCREs during hematopoietic differentiation. After jointly clustering cCREs based on their esRP scores across human and mouse cell types, we found distinctive transcription factor binding motifs associated with each group of cCREs with similar patterns of regulatory activity; these associations are similar between human and mouse. Genetic variants associated with blood cell phenotypes (from the GWAS Catalog and the UK Biobank study) were highly and specifically enriched in the catalog of human VISION cCREs, indicating the utility of our cCRE predictions for understanding the impact of noncoding genetic variants on both physiologic and pathologic blood cell-related traits. The cross-species joint modeling enabled a comparison of cCREs that revealed both conserved and lineage-specific patterns of epigenetic evolution, even in the absence of genomic sequence alignment. The VISION project resources are accessible through a suite of tools and browsers at http://usevision.org. Overall design: This paper reports an integrative analysis of genome-wide epigenetic data sets from progenitor and mature blood cell types jointly from human and mouse. A total of 404 data sets were used as input data for the integrative analysis, of which 216 were from human cell types and 188 were from mouse cell types. Most of the data used have been previously published. Many but not all of the previously published datasets have biological replicates. We also generated new data for multiple features in specific cell types to have more complete coverage of the epigenetic landscape across the cell types. For ChIP-seq of CTCF, we are submitting replicate data from G1E-ER4+E2 cells and one determination in G1E cells, along with the no-antibody "input" control files for each cell line.