Controlling for covariates and adjusting for multiple tests, the main effects of NR3C1 SNPs and intervention status on the intercept and slope of alcohol abuse were not significant. However, in conditional models that included the interaction terms between SNPs and intervention condition, one SNP (rs12655166) significantly moderated the effects of intervention among AAs, but not among EAs (see Figure 1). Results were the same for additive (see Table 2) and dominance models (see Supplementary Table 1) for this SNP. Intervention effects on the linear slope were stronger among AAs carrying C alleles of the SNP than those not carrying this minor allele (B = −0.97, SE = 0.24, p < .0001; for dominance effect, B = −1.02, SE = 0.27, p < .0001). Therefore, among AAs, for every 1 year increase, the intervention effect on the odds of not having any alcohol abuse relative to any abuse (i.e., preventing alcohol abuse) was 1.63 times stronger (e0.97 - 1) with each additional copy of this minor allele.
