Our primary association analyses of case-control status use logistic regression models which include age and gender as covariates. A standard series of regression equations to test main effects and gene – environment interaction were used. After the initial set of univariate logistic regression models for each covariate, we examined the main effects of each SNP and environmental risk factors jointly. Equation 1Gene&Environment Effect on Phenotype ln(p1−p)=α+β1G+β2h Equation 2Gene×Environment Interationn ln(p1−p)=α+β1G+β2h+β3hG where p is the probability of being a case, α is the intercept, h is parent monitoring coded in quartiles, and G represents genotype. Equation 1 provides adjusted estimates of risk due to genotype and environmental exposure; Equation 2 provides a test of genotype by environmental exposure interaction. Age and gender was also included as covariates in the models. For each SNP, we examined whether there is a statistically significant multiplicative interaction between parent monitoring risk levels and the risk gene represented by the SNP, as described in Regression Equation 2 in the above diagram. The interactions between specific risk and protective genotypes and lowest quartile parenting monitoring were tested.
