Consistent with our simulation results, mash identified many more significant effects than either mash-bmalite or ash. To avoid double-counting eQTLs in the same gene that are in LD with one another, we assessed the significance of only the “top” SNP in each gene (the SNP with the largest univariate |Z|-statistic across tissues). Thus, our results are based on 16,069 candidate eQTLs, each with effect estimates in 44 tissues, for a total of 44 × 16,069 = 707,036 effects.
