A limitation of our study is the sample size, which is small compared to case/control mega-GWASs that pool data across many different studies each having only limited phenotype information. In particular, future recruitment of additional methadone- and morphine-treated patients will be needed in order to study large numbers of CC homozygotes. Generally, pharmacogenomics GWASs tend to have many fewer subjects than studies of disease risk, because it is challenging to recruit and clinically characterize informative subjects, although the observed effect sizes are often greater in pharmacogenomics studies.69 Our GWAS is larger than all previously published opioid dose GWASs of which we are aware,41, 45, 46, 70 and our GS finding reinforces that these earlier GWASs were likely successful in identifying real signals despite modest sample sizes. Thus, the larger present sample would seem to be sufficient, especially considering the validation of rs73568641 in an independent sample of morphine-treated patients. Evaluation of rs73568641 using clinically documented morphine dose data also helps to compensate for another limitation of our study, which is that the data on usual daily methadone dose was collected
