The patterns of results should be different for “trans-eQTLs”, i.e., SNPs correlated with expression of a gene beyond 1 Mb of its genomic location. Trans-eQTLs incur a greater penalty for multiple testing, require greater power for detection, and are thus more susceptible to false positives and less likely to replicate than cis-eQTL. Nevertheless, the data supported 45,453 significant trans-eQTL at FDR ≤ 5%, of which 20,288 were also cis- eQTL SNPs for local genes, and 34% predicted expression of more than one distant gene. The proportion of trans eQTL in CMC that replicate in HBCC is 18.6% (both FDR ≤ 5%). The proportion of HBCC trans eQTL that replicate in CMC is 29.7%. Enrichment of trans-eQTLs with brain enhancers was not observed (data not shown), though enrichment in genic regions and depletion in intergenic regions was observed, particularly when restricting to trans eQTL ≥ 10 Mb from the gene location. We used similar techniques to derive isoform expression quantitative trait loci (isoQTLs).
