that tissue or cell line. To assess if this overlap is higher than expected by chance, we generated 1,000 sets of random SNPs matched with the index cis- and trans-eSNPs, in terms of allele frequency, gene density, distance from TSS, and density of tagSNPs arising from genomic variability of linkage disequlibrium. Z scores were estimated as: Z=observed−meannullSDnull Where observed is the number of index eSNPs that lie within enhancers, and meannull and SDnull are the mean and standard deviation of the null distribution of overlap, as estimated using the sets of permuted SNPs.
