We grouped the 21 counties of Iceland into 11 regions, as previously described [9] (Figure 1). From the entire set of 35,457 individuals, we selected a subset of 14,313 individuals with majority ancestry from one of the 11 regions, based on having at least 16 of 32 ancestors (five generations back) from that region (Table 1a). The goal of this scheme was to choose a set of samples reflecting the population structure of Iceland prior to the large-scale migration that resulted from industrialization and urbanization during the past century. From this set of 14,313 individuals we selected a further subset of 885 individuals—with at most 100 individuals from each region—that were unrelated at a meiotic distance of four generations. Of the 885 individuals, 8 were removed as genetic outliers when we ran PCA [17]; Table 1b and subsequent analyses are based on the remaining 877 individuals. The size limit of 100 individuals was used to ensure a relatively even representation of regions for analyses that are sensitive to varying sample sizes from subpopulations. We note that region 1, which contains the capital city of Reykjavik, was heavily underrepresented as it had a small population prior to urbanization.
