To illustrate how this approach might be used in practice to carry out a GWAS, we used the off-target exome data to compute association statistics at 103,977 SNPs across the genome using simulated phenotypes starting from the genotype calls from the arrays (Online Methods). We observed similar association statistics when imputed dosages were used as compared to SNP arrays under both null (phenotype uncorrelated to the genotype) and true nonzero effect sizes (Figure 2, Supplementary Figure 4,5,6, Supplementary Table 5), indicating that our approach is robust to false positives while accurately recovering the association signal when present. In addition, we also performed a case-control scan in which the AUT samples were treated as “controls” and SCZ as “cases”. After adjusting for differences in genetic ancestry between SCZ and AUT samples, we observed no genome-wide significant association, thus further emphasizing the robustness of our approach (Supplementary Note, Supplementary Figure 7). To assess the power of detecting true positives, in addition to simulated phenotypes, we also carried out a case-control study comparing HIV-1 controllers (61) and progressors (23) from the IHCS data
