SNPs exhibiting significant or near significant association with theta ERO measurements for both sample sets were further tested for association with diagnoses of alcohol dependence (DSM-IV). For the case-control data set, these analyses were performed in PLINK using logistic regression under an additive model, with age, sex, and the first and second PCs as covariates. For the family sample, a liability threshold model with an ascertainment correction was implemented in SOLAR, also assuming an additive mode of gene action. Markers that produced significant results under the additive model for either sample were additionally tested for association under both recessive and dominant genetic models (i.e., pairwise comparison of frequencies for either the minor or major homozygotes against the other two genotype classes). Haplotype-based tests were also carried out for the HTR7 region in the GWAS data set using PLINK, including single omnibus analyses for H–1 degrees of freedom that jointly tests all H haplotypes by comparing the alternate (each haplotype having a unique effect) versus the null (no haplotypes having any different effect) [Purcell et al., 2007b].
