eQTL studies can locate SNPs within a given GWAS locus that influence target transcript levels, but caution must be taken when interpreting results. First, eQTLs, like enhancers, are cell-type specific. Thus, the effect of a SNP on transcription may only occur in disease-relevant cell types [71,72]. Second, the SNP associated with transcript levels may not be the causal SNP: SNPs in LD with the eQTL SNP may be driving the association. Third, the results are correlative and may reflect indirect associations between SNPs and genes. Fourth, the effects on gene expression must be robust in order to be identified over the confounding effects of the genetic background. This poses a challenge for detecting functional variants that have modest effects, as has been proposed for most enhancer variants [7,33,73,74]. Fifth, eQTL analyses rarely consider the combinatorial effects of multiple SNPs at a given locus. Last, because eQTL studies are typically performed on healthy individuals, the impact of the SNP on the quantitative trait may differ in response to disease-specific stimuli. This was observed in a survey of enhancer SNPs associated with
