We applied the same methods to predict T2D in South Asian target samples from the UK Biobank (row 5 of Table 1). We used publicly available European summary statistics from DIAGRAM (12,171 cases and 56,862 controls; effective sample size = 40,101) as European training data, South Asian summary statistics from SAT2D (Kooner et al., 2011) (5,561 cases and 14,458 controls; effective sample size = 16,065) as South Asian training data, and UK Biobank genotypes and phenotypes (272 cases and 1,484 controls; effective sample size = 919) as South Asian validation data (see Methods).
