For trans-eQTL signals, we achieved low replication rates. The best result was for the Heinzen et al.14 data set, where 11.2% of trans-eQTL signals were replicated at a 5% type 1 error rate. Thus, while we did find evidence for the existence of true trans-eQTLs, we conclude that many of our trans-eQTL signals are likely to be false positives and therefore focus our attention here on our cis-eQTL signals. We expect that reliable trans-eQTL analysis will require the accrual of larger data sets.
