Figure 1 illustrates the result of our enrichment analysis applied to the WTCCC study of Bipolar Disorder. First, we took the top SNP associations (defined as p < 0.001; n=935) and asked whether there was an excess of cis mQTLs relative to null expectation. For the definition of mQTLs, we used the threshold of p < 0.01; a subsequent analysis using a more stringent p < 0.001 yielded similar results (so that the enrichment results appear robust to the definition of mQTL). To generate the null distribution, we conducted simulation studies, as previously described 6. See Materials and Methods for details of the simulation procedure. At the given mQTL threshold, we identified 132 such mQTLs among the top Bipolar Disorder associated SNPs (shown in Figure 1 as a black dot). From the simulated datasets (N=1000) each comprising of minor allele frequency-matched SNPs, the expected cis mQTL count was found to be 108.9 with standard deviation of 9.7. The enrichment p value is 0.01.
