To quantify the degree of concordance of effect size across populations, for those SNP-probe associations significant in multiple populations, we asked whether the SNP exerts the same effect size in each of the populations, as quantified by expression level fold-change differences between homozygote genotype categories in each of the populations. We observe that the effect size (fold difference between homozygotes of the two different genotypic states of a SNP) is shared between any two populations when the association is also shared (Figure 2), and furthermore, larger effect sizes were slightly more likely to be shared (Figure S9). In addition, for SNP-probe pairs discovered in one population, if we consider the p-value distribution in the other seven for the same SNP-probe pairs, we observe extensive enrichment of low p-values as indicated by the fraction of expected true positives pi1 (Figure S10), indicating that our threshold-based estimates of across-population cis-eQTL sharing are underestimates. This result, paired with the result that effect sizes (fold-change) are similar among populations, suggests that the driving force behind the discovery of an eQTL in one population but not another is mainly due to allele frequency differences and not due to differences in absolute effect size.
