To determine the level of (cis) mQTL (or eQTL or joint eQTL-mQTL) enrichment in the set of top susceptibility variants (defined in terms of p-value threshold) from a GWAS, we generated 1000 sets, randomly drawn without replacement from a control set (e.g., Affymetrix GeneChip 500K Mapping Array Set, the study platform, for the WTCCC GWAS), which match the minor allele frequency distribution of the SNP set of interest. For each random set, the (cis) mQTL (or eQTL or joint eQTL-mQTL) count was determined, yielding the distribution under the null. The observed (cis) mQTL (or eQTL or joint eQTL-mQTL) count enabled us to calculate an empirical p-value for the level of enrichment by considering the proportion of random sets with mQTL (or eQTL or joint eQTL-mQTL) count that matches or exceeds the observed count. We tested the dependence of our enrichment results on the definition of cis by doing these analyses when cis is defined as within 1 Mb of the gene.
