GWAS SNP enrichment analysis

All SNP‐disease/trait associations were taken from the NHGRI‐EBI GWAS Catalog Welter et al, 2014) (http://www.ebi.ac.uk/gwas/) downloaded on April 18, 2020. SNPs were filtered to retain only those located in the non‐coding genome (i.e., intronic and intergenic) (n = 100,108). We further obtained a list of SNPs associated with cancer (EFO_0000311) by GWAS and from this selected all traits corresponding to the types of cancer cell lines used in this studys (Table EV7, n = 2,312, hereafter referred to as cancer‐associated SNPs). For comparison of the overlap of cancer‐associated SNPs and enhancers to the overlap of SNPs from the 1000 Genomes project (Phase 3v5) and enhancers, we randomly sampled 1,000 matched SNP sets from the EUR population (with the same size as the original cancer‐associated SNP set), using SNPsnap (Pers et al, 2014) with default parameters (except r ² = 0.8). SNPsnap matched the original SNPs based on MAF, gene density, distance to nearest genes, and number of sites within LD of 0.8 of the original site. SNPs were then overlaid with TT‐seq‐defined enhancer regions, and the P‐value was determined based on the empirical background distribution.