additive model is easy to interpret and therefore the most commonly used. However, note that a lack of association in the additive model does not imply no association between the SNP and the trait. The genetic association can be adjusted for other covariates by adding them in the regression equation, or by modeling the residuals from the regression model that includes only the covariates. Adjustment should be done only for covariates that are significantly associated with the trait to avoid unnecessary loss of power. If the trait does not follow a normal distribution, the P-value can be computed using permutation methods, but the computational burden can be very high. Also, permutation methods usually tend to be less powerful [66].
