To extend our analysis to identify loci associated with asthma, we used data from the GABRIEL meta-analysis (31). We downloaded the genome-wide association P-values from the GABRIEL website (http://www.cng.fr/gabriel/results.html). We used the random-effects pooled P-values and excluded any SNPs that were not analyzed in at least 25 of the 36 studies, resulting in the inclusion of 526 351 SNPs. We calculated the median P-value for the set of candidate genes identified in our lymphocyte count GWAS, and assessed the significance by permutation testing, using the remaining genes (using the same approach described above). We again allowed for proxy SNPs when the paired (eQTL) SNP was not present in the GABRIEL data (Supplementary Material, Fig. S7B).
