In a first step, we used LDSC59 to estimate genetic correlations between the phenotypes in Supplementary Table 8. For phenotypes with multiple single-trait input GWAS files, we used the version with the largest Total N. This restriction leaves 53 single-trait input GWAS files, each of which is associated with a distinct phenotype. Because there may be sample overlap between the meta-analysed summary statistics, we used GWAS-equivalent sample sizes as reported by MTAG when estimating genetic correlations. (This was the case for Age First Birth, Number Ever Born (men), Number Ever Born (women), and Asthma/Eczema/Rhinitis. For the first three phenotypes, we meta-analysed the publicly available summary statistics from Barban et al.73, which included the first release of UKB, with UKB full release. Similarly, for Asthma/Eczema/Rhinitis, we meta-analysed publicly available summary statistics from Ferreira et al.49, which included the first release of UKB, with UKB full release.) The set of pairwise genetic correlations is reported in Supplementary Table 9.
