Testing gene-treatment interactions in pharmacogenetic studies

Drug-related side effects are one of the leading causes of death and illness in the developed world. Finding genes that modify drug response has the potential to significantly improve drug delivery, by identifying both individuals that can benefit from therapy and those at increased risk of harm. We present a simple approach to testing gene-by-treatment interactions in case-control pharmacogenetic studies. The approach utilizes a retrospective model that seeks to increase power through a Hardy-Weinberg equilibrium assumption among the controls, but does not assume that the event of interest is rare in the target population. We conduct extensive simulations and find that the approach shows similar or improved power, compared to standard methods, in all cases considered. We present methods for both autosomal and X-linked markers and show how the methods can be easily implemented using standard statistical software.