P value for that SNP will be set to 1.Select the SNP with minimum conditional P value that is lower than the cutoff P value. However, if adding the new SNP causes new colinearity problems between any of the selected SNPs and the others, we drop the new SNP and repeat this process.Fit all the selected SNPs jointly in a model and drop the SNP with the largest P value that is greater than the cutoff P value.Repeat processes (2), (3) and (4) until no SNPs can be added or removed from the model. A multiple regression analysis with model selection, such as that presented above, might suffer from over-fitting of effects, because the residual variance decreases as more and more SNPs are included in the model, such that the false positive rate for the inclusion of new SNPs in the model would be inflated. In practice, we keep the residual variance constant to the phenotypic variance, even if we added significant SNPs into the model, which may be too conservative and may therefore result in a loss of power to detect additional associated variants but has the benefit of keeping the false positive rate in the conditional and joint
