The UHS heroin addiction cases were compared to population controls assembled together from six study cohorts in the database of Genotypes and Phenotypes (dbGaP), as previously described (Hancock et al., in press). Exclusions were made for DSM-IV opioid dependence, where the data were available in the two study cohorts ascertained for addiction. Controls from the other four cohorts were ascertained for phenotypes unrelated to addiction, where we expected minimal phenotype misclassification for heroin addiction.
