Genetic Structure of the World's First Farmers Jul 26, 2018 18:33:39 GMT
Post by Admin on Jul 26, 2018 18:33:39 GMT
Population Genomic Analysis
We investigated the genome-wide relationships of our samples to modern and other ancient humans by performing principal-component analysis (PCA) defined by a subset of Eurasian populations from the Human Origins dataset [8, 9] (Figure 1B). Our Romanian genomes were projected onto the first two PCA axes, together with a broad panel of ancient individuals (see Data S1 for sample details), augmented by two newly sequenced Spanish Mesolithic hunter-gatherer genomes (Chan_Meso, an Early Mesolithic dated at 9,131 ± 124 cal BP (calibrated age before present), and Canes1_Meso, a Late Mesolithic dated at 7,115 ± 130 cal BP) to provide better coverage for this important period (Table 1 and STAR Methods, Method Details). The three Romanian Mesolithic genomes clustered together with other Mesolithic samples, including the two new Spanish ones, and close to modern Northern European populations (Figure 1B). The Romanian Eneolithic genome GB1_Eneo, on the other hand, was placed in a different region of the plot, located in a unique position between European Mesolithic hunter-gatherers and Early Neolithic farmers on PC2. We confirmed the intermediate nature of this genome by estimating ancestral components using the clustering algorithm ADMIXTURE. Whereas the Romanian Mesolithic hunter-gatherers had a single major ancestral component shared with other WHGs, the Romanian Eneolithic sample was characterized by a mixture between this WHG component and a component maximized in Neolithic farmers (Figures 1C and S3B). Other admixed European Neolithic and Eneolithic/Chalcolithic farmers had at most 20%–30% WHG ancestry in ADMIXTURE analysis, and the Romanian Eneolithic is the only individual who is genetically predominantly Mesolithic (61.7%, 95% confidence interval [CI] 59.9%–63.4%) with a minority contribution from the Neolithic. We note that Gok2, a Swedish Neolithic sample, was originally estimated to have 77% hunter-gatherer ancestry , but in our analysis it has a much lower percentage (27.2%, 95% CI 25.1%–29.4%), in line with other recent analyses . Although GB1_Eneo is chronologically close to the beginning of the Bronze Age, we did not find the green component (Figure 1C) characteristic of individuals from the Yamnaya culture, showing that the large hunter-gatherer component in this Eneolithic individual is unlikely to be due to admixture with incoming steppe pastoralists (Table 2).
Key D Statistics of the Form D(A,B; X,Y)
A B X Y D Z Score Loci
The Romanian Samples Form a Clade with Each Other (|Z| < 3; the Most Positive and Negative Statistics Are Shown for Each Comparison)
Mbuti Natufian SC1_Mes oSC2_Meso 0.03 1.635 26,749
Mbuti ElMiron SC1_Meso SC2_Meso −0.0281 −1.921 44,549
Mbuti GoyetQ116-1 SC1_Meso OC1_Meso 0.0272 1.896 41,197
Mbuti SC2_Meso SC1_Meso OC1_Meso −0.0283 −2.036 57,452
Mbuti Kostenki14 OC1_Meso SC2_Meso 0 .0231.964 79,553
Mbuti Armenia_EBA OC1_Meso SC2_Meso 0.0188 −2.192 77,923
There Is Anatolian-Farmer-Related Admixture in the Romanian Eneolithic Sample as Compared to the Romanian Mesolithic Samples
We formalized these inferences by computing outgroup f3 in the form f3(ancient1, ancient2, Mbuti), thus estimating the amount of shared drift between pairs of ancient samples with respect to an African outgroup. Our three Romanian Mesolithic samples share the most drift with each other (Figure 2), followed by other WHGs, including our new Spanish samples. The genetic affinity among the Romanian Mesolithic samples was such that they form a clade to the exclusion of other ancient samples, as shown by D statistics of the form D(Romanian Mesolithic 1, Romanian Mesolithic 2, other_ancient, Mbuti) (Table 2). Interestingly, this was not the case for the Spanish Mesolithic samples, as Chan is somewhat divergent from Canes1 and La Braña (Figure S3A and Table 2), highlighting the genetic diversity of European Mesolithic hunter-gatherer lineages. The Romanian Eneolithic individual, on the other hand, once again showed a mix of affinities. Based on outgroup f3, the genomes that shared the most drift with this Eneolithic sample are WHGs, in line with the large amount of that ancestry detected in ADMIXTURE. However, its affinity to Neolithic samples is also relatively high compared to the Romanian Mesolithic samples (Data S2). This conclusion is supported by D statistics of the form D(GB1_Eneo, Romanian HG, Anatolian Neolithic, Mbuti), which indicate some Near Eastern ancestry (Table 2). Our three Romanian hunter-gatherer samples are not direct representatives of the hunter-gatherer component in GB1_Eneo (Table 2); however, this might be due simply to the geographic distance between the sites, especially given the observed heterogeneity among Spanish Mesolithic hunter-gatherers.