Drift has been implicated as an explanation for population differences in polygenic scores among populations32, but others have reported that drift is insufficient to explain such differences33. Further, initial estimates of the strength of polygenic selection on height in European ancestry populations33,37 have recently been greatly reduced30,31, based on findings of uncorrected population stratification in summary statistics from the GIANT Consortium30,31. There is also disagreement about whether or not differences in average polygenic scores among populations might contribute to differences in phenotypic values among the same populations (which could also be due to environmental variation). Some have noted apparent positive correlations between average polygenic scores and phenotypes for BMI34, lupus35, and height as calculated using GIANT Consortium scores33,36,37. As described below, we include more data than used previously to address questions about potential correlations between worldwide height polygenic scores and height phenotypes.
