Finally, through resequencing and a widely used in vitro assay system [43], we found that of 14 tested genes, two genes harbored functional eQTNs in the proximal promoter and one gene harbored functional eQTNs in the 3′UTR. The success rate of 3 in 14 suggests that a substantial number of eQTLs, and by extension any complex traits that they may influence, can be functionally isolated using the scalable assay system that we employed or potentially higher-throughput assays [44]. We note that some truly functional variants will not be detectable in these assays, either from being tested out of their genomic context or having effect sizes below the limit of detection afforded by the number of replicates used (e.g. [45]), and that the actual fraction of eQTLs with promoter or 3′UTR functional variation may be substantially higher. Considering that replication was significantly greater for eQTLs near the ends of genes relative to those further away (Figure 3D), our functional analysis also strongly supports the use of SNP to gene distance as an important contributor to the prior probability that any given
