Two of the selected control cohorts were ascertained for addiction phenotypes: Collaborative Genetic Study of Nicotine Dependence (COGEND) and Study of Addiction: Genetics and Environment (SAGE). We excluded subjects with a positive report of opioid, marijuana, cocaine, or other drug dependence, as defined by the Diagnostic and Statistical Manual of Mental Disorders (4th ed.; DSM-IV). None of the other population control samples were ascertained for drug abuse or similar phenotypes and did not have drug use histories available.