Let Nj and pj denote total allele count and observed allele frequency in the Icelandic population, Nj 1 and pj 1 denote total allele count and observed allele frequency in ancestral population 1, and similarly Nj 2 and pj 2 in ancestral population 2, for SNP j. Let MIXα denote a synthetic population consisting of a linear combination of proportions α and (1−α) from ancestral populations 1 and 2, respectively. Let pjα = α pj 1+(1−α) pj 2. We estimate the F ST between Iceland and MIXα aswith the subtracted terms in the numerator adjusting for the effects of sampling error (see Supp Note 10 of [34]). We note that linkage disequilibrium between SNPs may lead to suboptimal weighting, which will increase the variance but will not bias the estimate. We estimate F ST for different values of α (on an evenly spaced grid from 0 to 1) and infer the ancestry proportion α that minimizes F ST, as described previously [35],[36]. We compute the standard error of the ancestry estimate α via a bootstrap approach. We partition the set
