A prediction model that consists primarily of genetic variants has a maximum AUC constrained by the heritability of the trait, as well as the disease prevalence in a population (Wray et al., 2010). As heritability of a disease goes down and as prevalence goes up, the maximum AUC goes down (Wray et al., 2010). This stresses the importance of taking into account other factors contributing to the variability in AD for risk prediction, particularly since AD is a fairly prevalent disorder. Additional measures to increase power may include reducing heterogeneity by refining the phenotype used as the outcome in the association study (Bierut et al., 2010). Large-scale meta-analysis, along with expanded individual association studies for AD, may improve the detection of disease variants.
