Figure 2: Relative rare allele sharing between ancient and modern samples.

(a) The ratio of the numbers of rare alleles shared with modern Dutch and Spanish samples as a function of the allele count in the set of modern samples. Ancient sample codes (left-hand and middle sections) are defined in Table 1. Results from present-day British individuals (right hand panel) are averaged over 10 individuals from each subpopulation. Results from a Dutch and a Spanish individual are shown for comparison. Error bars are calculated from raw count statistics and using s.e. propagation (Methods section). (b) The relative fraction of rare alleles shared with modern Dutch compared with Spanish alleles, integrated up to allele count five in the modern samples. Iron Age and Anglo-Saxon samples mark the two extremes on this projection, while modern samples are spread between them, indicating mixed levels of Anglo-Saxon ancestry, which is on average higher in East England than in Wales and Scotland, with a large overlap. Two Early Anglo-Saxon samples from Oakington have been excluded from computing the average, indicated by empty circles, because they show evidence for being admixed (O3) or of non-immigrant ancestry (O4). One modern sample from Scotland is also excluded, indidated as empty circle because it is a clear outlier with respect to all other Scottish samples. Samples are shown with a random vertical offset for better clarity. Error bars (Methods section) for the modern samples are omitted here, but of the same order of magnitude as for the ancient samples. Data for this figure is available as Supplementary Data 1.