For each of the four SNPs, prior to formal analysis, we determined the genetic risk model, i.e. if the effect was more likely due to the presence of a specific allele (e.g., A vs. G for rs1229984) or a specific genotype (e.g., AA, AG, or GG), by evaluating the prevalence of binary phenotypes (AD, any AUD, and risk drinking) and means of count phenotypes (AD severity, AUD severity, and maxdrinks) in each genotype group. Preliminary results showed similar prevalences and means for the alcohol phenotypes based on the presence or absence of a specific allele. For example, for rs12299884, both genotype groups with protective allele A (AA and AG) had similar phenotype values that were lower than group GG (absence of allele A), which was termed the “high risk” group. Consequently, further analysis compared the “high risk” group to the “low risk” group (AA or AG for rs1229984). Similarly, for the other three SNPs, high risk groups were as follows: rs1229982, GG (TT/GT low risk); rs1159918, TT (GG/GT low risk); rs698, GG (AA/AG low risk).
