Two susceptibility genes, in linkage disequilibrium with alleles of the markers D1G31 and D5G23, have been identified for the disease in the simulated data set of Problem 1. Here we apply the MASC (marker association segregation chi-square) method to model the joint effect of these two genes, by testing two-locus models. The model we obtain, that is the most parsimonious and that best fits the data, corresponds to a direct involvement of the alleles D1G31-8 and D5G23-7, with a nonmultiplicative effect of the two alleles. This was indeed assumed in the true model used for simulating the data.