Genomic structural equation modeling (gSEM) was used to evaluate how the MDD-META phenotype relates to 15 previously published large-scale GWAS of mental health and psychiatric phenotypes (See Online Methods and Discussion). Exploratory factor analysis (EFA) was conducted simultaneously on all traits and supported three- (cumulative variance = 0.605) and four-factor models (cumulative variance = 0.624) where each factor contributed over 10% to the cumulative explained variance. Anorexia nervosa did not load onto any factor during EFA and was therefore excluded from confirmatory factor analyses (CFA). CFA did not converge on a four-factor model due to high correlation between two factors. CFA of the three-factor model produced modest fit (comparative fit index = 0.884, X2[83df]=10034.76, AIC=10819.05, standardized root mean square error (SRMR)=0.086; Figure 4, Supplementary File 3). Factor 1 generally represented internalizing phenotypes with major contributions from depressive symptoms (loading = 0.95 ± 0.03), anxiety symptoms (loading = 0.92 ± 0.03), and posttraumatic stress disorder (loading = 0.92 ± 0.04). Factor 2 represented externalizing phenotypes with major contributions from risky behavior (loading = 0.85 ± 0.03) and cannabis use disorder (loading
