To integrate our independent data sets, we assessed whether genetic variation around genes that show the best evidence for differences in expression levels between individuals with high or low lymphocyte count is also implicated as associated with variation in lymphocyte count. To do so, we calculated the median lymphocyte count GWAS P-value for an expanding subset of gene–SNP pairs, after ordering genes based on the evidence for differences in expression levels between individuals with high or low lymphocyte count. To test if the observed median P-value is significantly lower than expected by chance alone, given the number of genes in each subset, we randomly sampled the same number of genes from the remaining list of gene–SNP pairs and calculated the median GWAS P-value 10 000 times. The level of significance (i.e. permutation P-value) was determined by counting how many sampled sets of gene–SNP pairs had a lower median P-value than the observed set.
