Additive SNP genotypes were tested for association with heroin addiction, separately by ancestry, using logistic regression models in SAS® software (SAS Institute, Cary, North Carolina) adjusted for sex and principal component eigenvectors. In each ancestry group, 10 eigenvectors were generated using EIGENSTRAT (Price et al., 2006); among these, the three eigenvectors which together explained >90% of the variance in heroin addiction case/control status were included as covariates in our regression models to circumvent any potential bias due to population stratification.
