We first present results from PGSGWS across different ancestry groups. PGSGWS-METAFE yielded prediction accuracies greater than or equal to that of all other PGSGWS (Fig. 4a), partly reflecting sample size differences between ancestry-specific GWASs and also consistent with previous studies29. PGSGWS-EUR (based on 9,863 SNPs) was the second best of all PGSGWS across ancestry groups except in AFR. Indeed, PGSGWS-AFR (based on 453 SNPs) yielded an accuracy of 8.5% (s.e. 0.6%) in AFR individuals from UKB and PAGE; that is, significantly larger than the 5.9% (s.e. 0.6%) and 7.0% (s.e. 0.6%) achieved by PGSGWS-EUR in these two samples, respectively (Fig. 4a). PGSGWS-METAFE was the best of all PGSGWS in AFR participants with an accuracy \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${R}_{{\rm{GWS}}-{\rm{METAFE}}}^{2}$$\end{document}RGWS−METAFE2 = (12.3% + 9.9%)/2 = 10.8% (s.e. 0.5%) on average between UKB and PAGE (Fig. 4a). Across ancestry groups, the highest accuracy of PGSGWS-METAFE was observed in EUR participants (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${R}_{{\rm{GWS}}-{\rm{METAFE}}}^{2}$$\end{document}RGWS−METAFE2~40%; s.e. 0.6%) and the lowest in AFR participants from the UKB (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy}
