Genetic studies are frequently limited to samples of European ancestry, precluding opportunities to assess differential genetic effects across ancestry groups. Although this study included a diverse sample, the non-European groups were of modest sample size and we lacked sufficient power to directly test such effects. However, efforts to recruit diverse samples are increasing, and meta-analytic approaches will enable the current samples to be incorporated into larger analyses in the future. Results from such approaches raise issues regarding the incorporation of linkage disequilibrium in various secondary analyses; here, we elected to use European linkage disequilibrium for FUMA, as this is likely a conservative approach and is appropriate for the majority (>82%) of the sample, but other methods may be preferable. JEPEGMIX was designed to be robust to the inclusion of cosmopolitan samples.
