Fig 1.
Mesolithic samples and their genetic affinities.
(A) Map of the Mesolithic European samples used in this study. The pie charts show the model-based [18,19] estimates of genetic ancestry for each SHG individual. The map also displays the ice sheet covering Scandinavia 10,000 cal BP (most credible [solid line] and maximum extend [dashed line] following [10]). Newly sequenced individuals are shown with bold and italic site names. SF11 is excluded from this map due to its low coverage (0.1×). Additional European EHG and WHG individuals used in this study derive from sites outside this map. The map was plotted using the R package rworldmap [28]. (B) Magnified section of genetic similarity among ancient and modern day individuals using PCA, featuring only the Mesolithic European samples (see S6 Text for the full plot). Symbols representing newly sequenced individuals have a black contour line. (C) Allele sharing between the SHGs, Latvian Mesolithic hunter-gatherers (Zv) [29], and EHGs versus WHGs measured by the statistic f4(Chimp, SHG; EHG, WHG) calculated for the captured SNPs [20]. Error bars show two block-jackknife standard errors. Data shown in this figure can be found in S1 Data. BP, before present; cal, calibrated; Chimp, Chimpanzee; EHG, eastern hunter-gatherer; PCA, principal component analysis; SHG, Scandinavian hunter-gatherer; WHG, western hunter-gatherer; Zv, Latvian Mesolithic hunter-gatherer from Zvejnieki.
Table 1.
Information on the seven SHGs investigated in this study, including cal BP (corrected for the marine reservoir effect, given as a range of two standard deviations), average genome coverage, average mt coverage, mt and Y chromosome haplogroups, and contamination estimates based on the mt, the X-chromosome for males and the autosomes.
Fig 2.
Migration scenarios into postglacial Scandinavia.
Maps showing potential migration routes into Scandinavia. Scenario (a) shows a migration related to the Ahrensburgian tradition from the south (S1 Text). Scenarios (b), (c), and (d) show different possible routes into Scandinavia for the EHG ancestry. The scenarios are discussed in the text and the scenario most consistent with genetic data and stone tools is a combination of routes (a) and (b). All maps were plotted using the R package rworldmap [28]. EHG, eastern hunter-gatherer.
Fig 3.
Genetic diversity in prehistoric Europe.
(A) RoH for the six prehistoric humans that have been sequenced to >15× genome coverage, (Kotias is a hunter-gatherer from the Caucasus region [35], NE1 is an early Neolithic individual from modern-day Hungary [27], the other individuals are described in the text), compared to all modern-day non African individuals from the 1000 Genomes Project [32]. (B) LD decay for five prehistoric populations each represented by two individuals (eastern SHGs: SF [SF9 and SF12], western SHGs: Hum [Hum1 and Hum2], CHGs [35]: [Kotias and Satsurblia], WHGs [18,35] [Loschbour and Bichon], and early Neolithic Hungarians [27]: EN_Hungary [NE1 and NE6]). LD was scaled in each distance bin by using the LD for two modern populations [32] as 0 (TSI) and as 1 (PEL). LD was calculated from the covariance of derived allele frequencies of two haploid individuals per population (S7 Text). Error bars show two standard errors estimated during 100 bootstraps across SNP pairs. (C) Effective population size over time as inferred by PSMC’ [44] for four prehistoric humans with high genome coverage. The dashed lines show the effective population sizes for selected modern-day populations. All curves for prehistoric individuals were shifted along the x-axis according to their radiocarbon date. S4 Fig. shows 100 bootstrap replicates per individual. Data shown in this figure can be found in S1 Data. BP, before present; CHG, Caucasus hunter-gatherer; LD, linkage disequilibrium; PEL, modern-day Peruvian individual; PSMC’, pairwise sequentially Markovian coalescent; RoH, runs of homozygosity; SHG, Scandinavian hunter-gatherer; TSI, modern-day Tuscan individual; WHG, western hunter-gatherer.
Fig 4.
Adaptation to high-latitude environments.
(A) Plot of similarity between Mesolithic allele frequency and FIN allele frequency in contrast to difference to TSI allele frequency using the statistic Dsel. The figure shows all positive Z scores representing the number of standard deviations each SNP deviates from the mean. The green-highlighted SNPs are all located in the TMEM131 gene. The plot was made with qqman [49]. (B) Derived allele frequencies for three pigmentation-associated SNPs (SLC24A5, SLC45A2, which are associated with skin pigmentation, and OCA2/HERC2, which is associated with eye pigmentation). The dashed line connecting EHG and WHG represents potential allele frequencies if SHG were a linear combination of admixture between EHG and WHG. The solid horizontal line represents the derived allele frequency in SHG. The blue symbols representing SHGs were set on the average genome-wide WHG and EHG mixture proportion (on x-axis) across all SHGs, and the thick black line represents the minimum and maximum admixture proportions across all SHGs. Dashed horizontal lines represent modern European populations (CEU). The p-values were estimated from simulations of SHG allele frequencies based on their genome-wide ancestry proportions (S9 Text). Data shown in this figure can be found in S1 Data. CEU, Utah residents with Central European ancestry; EHG, eastern hunter-gatherer; FIN, modern-day Finnish individual; SHG, Scandinavian hunter-gatherer; TSI, modern-day Tuscan individual; WHG, western hunter-gatherer.