GWAS3D integrates multiple genome-wide experimental data to connect genetic variants with underlying gene regulation mechanism through high-dimensional regulatory interactions. We first collected and curated the experimental results of long-range interactions, for 16 different cell types, measured by high-throughput chromosome conformation capture technologies (5C, ChIA-PET and Hi-C) from the ENCODE project, Gene Expression Omnibus (GEO) database and published resources (Supplementary Table S1). We directly used 5C and ChIA-PET interactions in the database and processed the Hi-C interactions by the iterative correction and eigenvector decomposition (ICE) algorithm (25), which can largely reduce the false positives and biases. Some chromatin marks have been reported and validated as the active signals of enhancers, including histone modifications of H3K4me1 and H3K27ac, DHSs and E1A-binding protein p300 (26,27), we therefore extracted the related ChIP-Seq peaks for the above 16 cell types from ENCODE. We also collected ChIP-Seq data of CTCF-binding sites, which imply transcription repression and chromatin insulation. For predicted elements, we downloaded the ChromHMM genome-wide maps of chromatin state annotations for supported cell types and extracted the promoter, enhancer and insulator elements whose signals are
