As a complementary to MiXeR we estimated local genetic correlations (rg) of depression and the additional traits listed above using LAVA47. LAVA estimates local rg across 2,495 semi-independent genetic loci of approximately equal size (~1 Mb) in a two-step process: 1) univariate analysis: the observed h2SNP is estimated for each locus for each trait. 2) bivariate analysis: the local genetic covariance is estimated for each locus using the method of moments. Genetic loci included in the bivariate analysis were filtered according to their local h2SNP using a significance threshold of P<10−4, consistent with previous usage of LAVA47,91. Sample overlap was controlled using linkage disequilibrium score regression36. Significance testing was performed using simulation-based P-values and we used the false discovery rate (FDR) to adjust for multiple testing, reporting loci with FDR<0.05.
