Due to this highly significant enrichment of converging pairs of SNPs and its low estimated false-positive rate, we also ran an analysis where we had relaxed the FDR for trans-eQTLs to 0.50 (Table S7). Here we observed 18 pairs of SNPs that converge on the same genes, whereas in the 100 subsequent permutations we observed this only on average for 0.84 SNP-pairs (range [0, 5], 21 times more expected by chance, EVD p<1.0×10−16, implying a false-positive rate of 0.047, Table 2, Figure 4b).
