The pattern of SNPs classified into varying numbers of classes further supported the existence of clusters of SNPs corresponding to externalizing and internalizing genetic factors, especially between 3- and 4-class models. The AD+ASP class in the 3-class model split into AD or ASP classes in the 4-class model. However, almost no overlap (0.3% of SNPs) was observed between the AD+ASP class from the 3-class model, and the MD class identified in the 4-class model. SNPs in the MD class from the 3-class model mostly (83.4%) remained in the MD class in the 4-class model. This pattern of classification broadly reflects the hierarchical factor structure identified from factor analyses (Krueger 1999; Krueger et al. 2001; Krueger and Markon 2006) and twin studies (Kendler et al. 2003; Kendler et al. 2011a; Krueger et al. 2002). These findings suggest that the application of LPA to SNP association results across multiple disorders may be a promising approach to identify clusters of SNPs that are associated with related phenotypes. While the cross-tabulations of class memberships between 2- to 4-class models matched our expectation, some instability
