Haplotype frequencies were estimated using a Bayesian approach implemented with PHASE (Stephens & Donnelly, 2003). Linkage disequilibrium (LD) relationships of the 19 SNPs were examined using Haploview software (Barett et al. 2005) the PHASE v2.1.1 Software (Stephens et al., 2001). Haplotype blocks were defined using validated a-priori methods (Gabriel et al., 2002). In order to test the effects of CRHR1 haplotypes on the likelihood of meeting criteria for alcoholism, as well as its interaction with trauma exposure, logistic regression models were run for each haplotype block using R Statistical Software. These models included sex (to correct for sex imbalance between cases and controls) and trauma exposure as covariates, and each haplotype within the respective block (1 or 2), as well as the interaction between haplotype and traumatic stress (G × E). For all haplotype analyses the most frequent haplotype was used as the reference group.
