The availability of high-throughput genotyping and sequencing platforms has largely removed technological barriers in the mapping the genetic determinants of drug response in human populations, and the set of validated pharmacogenetic variants is gradually increasing. Like the search for disease-susceptibility variation, however, many of the loci identified to date represent the relatively low-hanging fruit with large phenotypic effects but relatively low predictive power. Yet to be discovered is the larger set of variants, each with considerably weaker phenotypic effects, which together can be used to predict drug response more reliably and identify potential targets for novel drug development. Finding these pharmacogenetic variants is particularly challenging because sample size is typically far too small (and thus statistically underpowered) to detect genetic variants with weak effects. Studies of the genetics of gene expression (also described as expression quantitative locus (eQTL) mapping or genetical genomics) represent a novel approach for identification of functional genetic variants that influence gene expression. In these studies, individual gene transcript abundance as measured from expression microarrays are considered as discrete quantitative traits for genetic mapping that are intermediate to clinical outcomes of interest. Early studies using these methods have demonstrated improved power to detect such regulatory variants and have facilitated mapping of disease-susceptibility variants. The potential use of this approach in the study of pharmacogenetics and for the identification of potentially modifiable drug targets is reviewed here.