Fig. 2

Illustration of the main steps of the simulation data generation process. The dark squares represent the SNP array, with two populations (marked with red descriptions). We group every five SNPs and simulate their interaction, result in one epistatic variable. For each epistatic variable, we introduce an effect size. Summing over the effects introduced by these epistatic variable, together with the effects introduced by population structure, we result in an continuous variable, which will further be transformed into binary phenotype