The MiXeR bivariate Gaussian mixture modeling revealed compelling genetic overlap between depression and other traits when considering both concordant and discordant variants, suggesting that psychiatric disorders and correlated traits are substantially more intertwined than indicated by their genetic correlations. The most drastic results were observed for educational attainment, showing that 99% of the depression risk variants are also influencing educational attainment and, vice-versa, 88% of educational attainment variants are influencing depression. This almost complete overlap in influencing variants, with a majority of discordant variants (58%), is consistent with overall negative genetic correlation between the two traits (rg=−0.23) and refines the understanding of their polygenic architecture appreciably. Although generally less pronounced, a similar picture was observed for the other mental health and behavioral traits examined. These results corroborate and extend previous results on partly older and smaller GWASs92. A combination of (altogether) only around 10,000-15,000 variants appear to explain 90% of the SNP heritability of the investigated major psychiatric disorders and correlated behavioral and cognitive traits. This narrows the search-space for risk variants to around 1% of the 106 quasi-independent SNPs
