The promise of pharmacogenomics. Genetic risk factors for disease lie along a spectrum of rare to common variants and weak to strong effect sizes. For genes conferring risk to disease, variant cannot be common and strong at the same time. If they were, these variants would already have been identified long ago through linkage studies. However, for possible genes associated with “secondary phenotypes”, these restrictions do not exist and there are various examples of common and strong modifier genes in other diseases. Do we find a similar situation in genetic risk factors associated with response to treatment with antiepileptic drugs?