single population. For the two African populations, where low-frequency diversity is greater and the populations more diverged (FST = 0.008), the difference between reference and target populations was more substantial, with mean r2 only rising above 0.7 when five copies of the minor allele were in the reference panel. In both cases, however, the cross-population accuracy was much better than that seen in Supplementary Table 10, indicating that cross-population loss of accuracy largely results from the incomplete sharing of low-frequency alleles between reference and target samples, rather than from differences in haplotype backgrounds.
