|
|
Post by Admin on Aug 28, 2021 4:31:56 GMT
Dynamic changes in genomic and social structures in third millennium BCE central Europe
Science Advances 25 Aug 2021:
Vol. 7, no. 35, eabi6941
DOI: 10.1126/sciadv.abi6941
Abstract
Europe’s prehistory oversaw dynamic and complex interactions of diverse societies, hitherto unexplored at detailed regional scales. Studying 271 human genomes dated ~4900 to 1600 BCE from the European heartland, Bohemia, we reveal unprecedented genetic changes and social processes. Major migrations preceded the arrival of “steppe” ancestry, and at ~2800 BCE, three genetically and culturally differentiated groups coexisted. Corded Ware appeared by 2900 BCE, were initially genetically diverse, did not derive all steppe ancestry from known Yamnaya, and assimilated females of diverse backgrounds. Both Corded Ware and Bell Beaker groups underwent dynamic changes, involving sharp reductions and complete replacements of Y-chromosomal diversity at ~2600 and ~2400 BCE, respectively, the latter accompanied by increased Neolithic-like ancestry. The Bronze Age saw new social organization emerge amid a ≥40% population turnover.
INTRODUCTION
Archaeogenetics has revealed two major population turnovers in Europe within the past 10,000 years (1–5). The first, beginning in the seventh millennium BCE, was associated with expanding Neolithic farming communities from Anatolia (6, 7). European Early Neolithic farmers were initially genetically distinct from preceding hunter-gatherers (HG) and almost indistinguishable from Anatolian farmers (8–10), however incorporated HG ancestry into their gene pools over ensuing millennia (3, 11–13).
The second major turnover occurred in the early third millennium BCE with individuals of the Corded Ware (CW) culture (3, 4, 8). Of note, in what follows, we use the co-occurrence of human skeletal remains and markers of archaeological cultures (e.g., grave goods and body orientation) to denote an association between individuals and an archaeological culture (e.g., “CW individuals”), although this may not reflect a unified social entity. The CW represents a major cultural shift in central, northern, and northeastern Europe, bringing changes in economy, ideology, and mortuary practices (14–22). CW individuals were shown to be genetically distinct from culturally pre-CW people, having ~75% of their ancestry similar to Yamnaya individuals from the Pontic-Caspian steppe (3, 4, 23–27). This Yamnaya-like “steppe” ancestry then spread rapidly throughout Europe, reaching Britain, Ireland, the Iberian Peninsula, the Balearic Islands, Sardinia, and Sicily before the end of the third millennium BCE (5, 28–32).
Despite the importance of the third millennium BCE, our genetic understanding is mainly built upon studies with pan-European sampling strategies, with little emphasis on regional, high-resolution temporal transects (3–5, 8). Consequently, many temporal and geographic sampling gaps remain, resulting in limited knowledge about the processes at the level of the societies and communities and how cultural groups interacted, influenced, and gave rise to one another. In addition, the use of small sample sizes to represent supra-regional archaeological phenomena, as well as the resulting oversimplified culture-historical interpretations, has drawn criticisms from archaeologists (21, 33–40).
Unresolved questions concern the genetic and geographic origins of CW and Bell Beaker (BB) individuals, their relationship to one another and to Yamnaya individuals, as well as the origin of Early Bronze Age (EBA) Únětice individuals. Although it has been proposed that CW formed from a male-biased westward migration of genetically Yamnaya-like people (23, 41–44), no overlap in Y-chromosomal lineages (with the exception of a few nondiagnostic I2) has been found between the predominantly R1a-carrying CW and mainly R1b-Z2103–carrying Yamnaya males. Steppe ancestry is also present in BB individuals (5); however, they predominantly carry R1b-P312, a Y-lineage not yet found among CW or Yamnaya males. Therefore, despite their sharing of steppe ancestry (3, 4) and substantial chronological overlap (45), it is currently not possible to directly link Yamnaya, CW, and BB groups as paternal genealogical sources for one another, particularly noteworthy in light of steppe ancestry’s suggested male-driven spread (23, 41–43) and the proposed patrilocal/patriarchal social kinship systems of these three societies (46–48).
Crucial to understanding the cultural, social, and genetic transitions in third millennium BCE Europe are densely settled regions that attest to the (co)existence of societies attributed to pre-CW [Baden and Globular Amphora (GAC)], CW, BB, and EBA Únětice. Currently, no such region has been systematically studied from the archaeogenetic perspective. Situated in the heart of Europe and tightly nestled around the important Elbe river, the fertile lowlands of Bohemia, the western part of today’s Czech Republic, witnessed many major supra-regional archaeological phenomena (table S1, Fig. 1, fig. S1, and the Supplementary Materials). Dense agrarian settlement of Bohemia began after ~5400 BCE (49, 50) with the arrival of early Neolithic farmers (Linearbandkeramik-LBK, later Stichbandkeramik-STK, and Lengyel). They were succeeded by manifold societies of the Eneolithic (~4400 to 2200 BCE), associated with more than a dozen archaeological cultural groups including Jordanów, Michelsberg, Funnelbeaker, Baden, Řivnáč, GAC, Early and Late CW, and BB (table S1) (50). The Eneolithic witnessed important innovations (metallurgy, the wheel, wagon and plough, fortified hillforts, and burial mounds) (51–53) and was succeeded by the globalized EBA Únětice culture, geographically centered around Bohemia.
|
|
|
|
Post by Admin on Aug 28, 2021 21:07:00 GMT
 Fig. 1 Temporal and geographic distribution of studied Neolithic, Eneolithic, and EBA individuals from Bohemia. (A) Map of Bohemia showing the locations of sampled sites (red, new; blue, previously published; table S2 and figs. S1 to S5). (B) Mean age of newly reported (n = 206) and published (n = 65) individuals from Bohemia. (C) Local chronology of archaeological cultures and time periods. Black triangles indicate external influences visible in the material culture. Red lines indicate qualitative degree of change in material culture. In addition to material and technological developments, ideological changes, as manifested through mortuary behavior, are also evident (54). Although relatively common during the Funnelbeaker period (~3800 to 3400 BCE, n = ~100 known graves in Bohemia) (55), regular graves almost disappear from the succeeding Baden, Řivnáč, and GAC periods (Middle Eneolithic, ~3500 to 2800 BCE, n = ~20 in Bohemia) (56). Single graves, but now with strict gender differentiation in body position and grave goods, reappeared in abundance with CW from ~2900 BCE (n = ~1500 in Bohemia) (50, 57) and continued with BB (n = ~600 in Bohemia) from ~2500 BCE (58), who developed and maintained important differences from the preceding CW. The EBA Únětice culture (59, 60) continued with single graves (n = ~4000 to 5000 in Bohemia), but now again without gender differentiation in body position. To better understand these transitions, we analyzed a high-resolution archaeogenetic time transect of 271 (206 newly reported and 65 previously published) individuals (Fig. 1, fig. S1, tables S2 to S4, and the Supplementary Materials) from the northern part of Bohemia. Through dense genetic sampling from geographically and temporally overlapping archaeological cultures, we aim to (i) address whether cultural changes in the Eneolithic and EBA of central Europe were driven by an influx of nonlocals, (ii) characterize the central European genetic diversity immediately prior the appearance of CW, (iii) date when individuals with Yamnaya-like steppe ancestry first appeared in central Europe and understand their genetic origin and social structure, (iv) characterize the nature and extent of biological exchange between the “locals” and “migrants” after the appearance of CW, and (v) identify social transformations linked to genetic and archaeological changes. RESULTS General sample overview We screened 261 prehistoric individuals (table S3) from 37 sites (table S2) for ancient human DNA preservation, of which 219 individuals were enriched for 1,233,013 ancestry informative sites in the human genome (“1240k capture panel”) (8). After enrichment, individuals with fewer than 30,000 sites covered (on 1240k) or signs of contamination were removed (n = 13), resulting in a dataset of 206 newly reported individuals. We combined our dataset with 65 previously published individuals (5, 61, 62) from Bohemia (with >30,000 covered sites on 1240k; table S4) and wider (table S5), thereby extending the total number of published Bohemian Neolithic and pre-CW Eneolithic individuals from 7 to 58 (fig. S2), CW individuals from 7 to 54 (fig. S3), BB individuals from 40 to 64 (fig. S4), and EBA individuals from 11 to 95 (fig. S5). Crucially, we substantially expand the sample size of individuals around the time of CW formation (~3200 to 2600 BCE, from n = 1 to n = 50; Fig. 1B), i.e., the last pre-CW (Baden, Řivnáč, and GAC, from n = 0 to n = 18) and the early CW (from n = 1 to n = 32) individuals, allowing us to directly study the origin of CW in central Europe, the nature of their migration, and social interactions with coexisting pre-CW people. First-degree relatives were excluded from allele frequency–based analyses [f statistics, qpWave, qpAdm, Distribution of Ancestry Tracts of Evolutionary Signals (DATES), and Y chromosome analyses; table S4 and see Materials and Methods]. We also report 140 new radiocarbon dates to aid in finer temporal resolution, allowing us to study the genetic changes between early and late phases of important third millennium BCE cultural groups (e.g., CW, BB, and Únětice; tables S4 and S6). Bohemia before Corded Ware (pre-CW, before ~2800 BCE) We first assessed the genome-wide data by projecting the ancient individuals from Bohemia onto the first two axes of a principal components analysis (PCA) constructed from 1141 modern-day West Eurasian individuals (table S7). In the resulting PCA plot (Fig. 2A), all (n = 58) pre-CW individuals from Bohemia plot between Anatolia_Neolithic and Western HG (WHG), in close proximity to published culturally pre-CW individuals from central Europe (3, 8, 11, 13, 25). This suggests an absence of steppe ancestry, which we formally confirmed using qpAdm modeling (table S8), revealing that pre-CW individuals from Bohemia can be largely modeled as two-way mixtures of Anatolia_Neolithic and WHG (Fig. 3A, tables S8 and S9, and fig. S6). The percentage of HG ancestry is positively correlated with time (Spearman’s rank correlation r = 0.39, P < 0.004), showing that the previously reported trend of increasing HG ancestry during the Neolithic also took place in Bohemia (3, 11). We found this HG ancestry increase to be best modeled as a two-stage linear process (Fig. 3A, table S8, and the Supplementary Materials), with an increase in HG ancestry during the fifth millennium BCE, followed by stasis (nonsignificant slope) thereafter (Fig. 3A and Supplementary Materials). |
|
|
|
Post by Admin on Aug 29, 2021 1:43:23 GMT
 Fig. 2 PCA of published and newly reported ancient individuals from Bohemia (n = 271). Data are displayed in four major time periods: (A) Pre-CW–Eneolithic, (B) CW, (C) BB, and (D) EBA. Modern-day West Eurasian individuals upon which principal components were calculated (n = 1141; table S7) are grayed out in the background with modern-day Czech and relevant ancients (projected) plotted as colored polygons for reference [labeled in (A), WHG, EHG, Latvia Bronze Age (BA), Yamnaya Samara/Kalmykia, and Anatolia Neolithic]. Individuals mentioned in the main text are labeled.  Fig. 3 WHG ancestry in pre-CW Bohemia. (A) The proportion of WHG ancestry through time in pre-CW individuals from Bohemia modeled as a two-stage linear process (table S8). The gray area indicates 95% confidence interval. (B) Left: Proportion of ancestry ascribable to Anatolia_Neolithic, Loschbour, and Körös_HG in pre-CW cultural groups from Bohemia in chronological order from bottom to top (sample size of each cultural group in brackets and P value of three-way qpAdm model indicated within orange bars; table S10). Right: Inferred dates (2 SE; table S11) of HG admixture (gray interval) relative to culture’s chronology (black interval). (C) Zoomed-in PCA showing (with the exception of TUC003) segregation between Bohemian GAC and Řivnáč individuals along with position of early CW females without steppe ancestry (green circles). Black dots represent previously published GAC individuals from Poland and Ukraine. To gain insight into the process(es) by which HG ancestry was incorporated into the gene pool of pre-CW individuals from Bohemia, we used qpAdm to model each pre-CW cultural group as a three-way mixture of Anatolia_Neolithic, Loschbour, and Körös_HG as well as DATES to estimate the introgression date of incorporated HG ancestry (Fig. 3B and tables S10 and S11). Under a scenario of population continuity with sequential incorporation of HG ancestry, the mean date of introgression, as indicated by the gray intervals in Fig. 3B (right), for succeeding cultures is expected to become more recent through time. Conversely, under population continuity without incorporation of further HG ancestry, the admixture date should be similar for successive cultural groups who have similar HG proportions. Our results indicate two cultural transitions for which either of these expectations is not met. First, although Bohemian Jordanów and Funnelbeaker have similar amounts of WHG ancestry [f4(Mbuti.DG, WHG; Jordanów, Funnelbeaker) ~0; z score, 0.96], the estimated date of WHG introgression for Funnelbeaker is significantly earlier (5079 to 4748 BCE) than for Jordanów (4636 to 4310 BCE) (Fig. 3B and table S11), consistent with Bohemian Funnelbeaker individuals being derived from a different population (whose HG ancestry was incorporated further back in time), which superseded the Jordanów population in Bohemia. This transition between Jordanów and Funnelbeaker is corroborated by three additional observations. First, an f4 statistic of the form f4(Mbuti.DG, Bohemia-Funnelbeaker; Bohemia-Jordanów, Germany-Funnelbeaker) is positive (z score, 3.14), revealing significantly greater genetic affinity of Bohemian Funnelbeaker to Funnelbeaker (Baalberge and Salzmünde) individuals from Saxony-Anhalt than to the preceding local Jordanów individuals. Conversely, f4(Mbuti.DG, Bohemia-Jordanów; Germany-Funnelbeaker, Bohemia-Funnelbeaker) is consistent with 0 (z score, 1.03), suggesting phylogenetic cladality between Bohemian and German Funnelbeaker with respect to Bohemia-Jordanów individuals. Second, Bohemia-Jordanów individuals can be modeled as a two-way mixture of Anatolia_Neolithic and Körös_HG but not Anatolia_Neolithic and Loschbour, while the opposite is true for Bohemia-Funnelbeaker (table S12). This suggests different affinities, in addition to the different introgression dates, of the HG ancestries in Bohemian Jordanów and Funnelbeaker cultural groups. Third, qpWave does not support cladality between Bohemian Jordanów and Funnelbeaker (P = 0.00679), while cladality between Bohemian and German Funnelbeaker cannot be rejected (P = 0.88; table S13). Together, these results indicate a largely (significantly more than 50%) nonlocal genetic origin of Bohemian Funnelbeaker individuals. The second such case can be seen in the Řivnáč to GAC cultural transition. GAC individuals carry the most HG ancestry among pre-CW cultural groups from Bohemia (25.7%, ±1.4), significantly more than Řivnáč individuals [f4(Mbuti.DG, WHG; Řivnáč, GAC) >> 0; z score, 4.46]. However, the estimated date of HG admixture in GAC is not later than in Řivnáč individuals (Fig. 3B and table S11), suggesting that GAC individuals do not descend from a recent mixture of Řivnáč and an HG source but instead constituted a recent, nonlocal incursion in Bohemia from a region that received more HG gene flow [e.g., Poland (13, 63)], in accordance with interpretations of archaeological evidence (56). A distinct genetic origin for Řivnáč and GAC individuals is further supported by PCA and qpAdm modeling. From PCA, we find that with the exception of TUC003, Řivnáč and GAC individuals form distinct clouds (Fig. 3C). This is confirmed by qpAdm modeling where GAC individuals can be modeled as a mixture of Anatolia_Neolithic and Loschbour but not Anatolia_Neolithic and Körös_HG, while the opposite is true for Řivnáč individuals (table S14). Consequently, Řivnáč and GAC individuals are distinguishable based on the amount and source of HG ancestry, suggesting that Bohemia was inhabited by genetically differentiated groups of Řivnáč and GAC individuals at the time of CW appearance. The Řivnáč outlier (TUC003) also raises the interesting possibility of an individual born into a GAC but buried in a Řivnáč cultural context. Among the 16 Řivnáč and GAC individuals who are contemporaneous with or postdate the appearance of CW in Bohemia (Fig. 1B), we find no detectable traces of steppe ancestry (Fig. 2A and table S8), suggesting that biological exchange from CW/Yamnaya into culturally pre-CW people (e.g., Řivnáč and GAC) was low, possibly nonexistent. Steppe ancestry coappears with CW individuals in early third millennium BCE Bohemia. |
|
|
|
Post by Admin on Aug 29, 2021 22:11:12 GMT
Corded Ware
We report genomic data from the earliest CW individuals to date, including STD003 (northwestern Bohemia, 3010 to 2889 calibrated (cal) BCE), VLI076 (central Bohemia, 3018 to 2901 cal BCE), OBR003 (central Bohemia, 2911 to 2875 cal BCE), and PNL001 (eastern Bohemia, 2914 to 2879 cal BCE), showing that CW was widespread across Bohemia by 2900 BCE. The early radiocarbon dates are also supported by these individuals’ genetic profiles, who occupy the most extreme positions on PC2 (Fig. 2B), as expected under a scenario of the earliest CW being migrants from the east who mixed with locals, resulting in intermediate PC2 positions in later generations.
To explore the formation of the Bohemian CW gene pool, we grouped CW individuals with steppe ancestry and mean age > 2600 BCE (n = 27) into a Bohemia_CW_Early group and the rest (n = 21) into Bohemia_CW_Late (table S4). We found poor statistical support (P < 0.005) for modeling Bohemia_CW_Early as a two-way mixture of any known Yamnaya source and any local Bohemian or nonlocal pre-CW source from Poland, Ukraine, Hungary, or Germany (table S15). When using distal sources as proxies for the Neolithic ancestry (Anatolia_Neolithic and a range of HG sources), we found no strong support (P < 0.05) for all but one of the three-way distal models (table S16). However, this one statistically supported model results in a previously unobserved ratio of Neolithic ancestry in Europe (i.e., a Neolithic population of ~1:1 ratio of Anatolia_Neolithic:Sweden_Motala_HG). In addition, when modeling early CW individually as “standard” three-way mixtures of Anatolia_Neolithic, WHG, and Yamnaya_Samara (3), we find that in 37% (10 of 27) of cases, the model lacks strong support (P < 0.05 or infeasible; fig. S6 and table S9).
To explore why two-way proximal models between any Yamnaya and a European Neolithic source are insufficient in explaining Bohemia_CW_Early genetic diversity, we tried adding a third source to obtain better model fits. We find that when either one of Latvia_MN, Ukraine_Neolithic, or PittedWare is added as a source, almost all (280 of 285) model fits (P values) improve and most of them by several orders of magnitude (table S17). While all (n = 95) two-way proximal models lack strong support (P < 0.05; table S17), the addition of either Latvia_MN (57 of 95 supported models), Ukraine_Neolithic (53 of 95 supported models), or PittedWare (32 of 95 supported models) to the sources drastically increases the number of supported models (table S17). These results show the presence of excess Latvia_MN/Ukraine_Neolithic/PittedWare-like ancestry in Bohemia_CW_Early relative to all known Yamnaya and central European Neolithic groups. Our models suggest that this ancestry accounts for ~5 to 15% of the Bohemia_CW_Early gene pool (table S17). Increases in model fits with either of these third sources are also observed when modeling Bohemia_CW_Late and Germany_Corded_Ware, suggesting this ancestry to be present also in later central European CW (tables S18 and S19) and is consistent with allele sharing f4 statistics, which show that CW groups share more alleles with ancient northeast European groups than do Yamnaya (tables S20 and S21).
We provide the first genomic data from CW individuals without steppe ancestry, thereby elucidating the social processes of interaction between CW and pre-CW people. Observing only females (four of four) among early CW individuals without steppe ancestry (Figs. 2B and 3C) suggests that the process of assimilating pre-CW people into early CW society was female-biased. Two of these females (STD003 and VLI008) plot in close PCA space to GAC individuals from Bohemia and Poland (Fig. 3C). When grouped together, we find that STD003+VLI008 share more genetic affinity with Bohemian GAC than with Bohemian Řivnáč [f4(Mbuti.DG, STD003+VLI008; Bohemia-GAC, Bohemia-Řivnáč) < 0; z score, −2.32]. These two females are not genetically closer to Bohemian compared to Polish GAC individuals [f4(Mbuti.DG, STD003+VLI008; Bohemia-GAC, Poland-GAC) ~ 0; z = 0.5], meaning that a nonlocal, (north)eastern origin (e.g., Poland) cannot be ruled out. In addition, VLI009 and VLI079 fall outside of the sampled Bohemian Middle Eneolithic (Baden, Řivnáč, and GAC) genetic variation in PCA, carrying significantly more HG ancestry (Fig. 3C and table S22), suggesting that a large proportion (50%, or higher when including STD003/VLI008) of the genetically pre-CW females of the early CW society originated from outside Bohemia.
We find that Bohemia_CW_Late carries significantly more pre-CW–Eneolithic–like ancestry compared to Bohemia_CW_Early (table S23); however, this signal is lost when early CW females without steppe ancestry are included (table S24). This additional pre-CW–Eneolithic–like ancestry in Bohemia_CW_Late (relative to Bohemia_CW_Early) is poorly modeled as coming from local sources (table S25), suggesting nonlocal genetic influences on the Bohemian CW gene pool through time. This is consistent with the genetically pre-CW females originating from outside of Bohemia and is supported by the finding that Bohemia_CW_Early (including females without steppe ancestry) and Bohemia_CW_Late are not cladal in qpWave analysis (table S26), despite having similar amounts of pre-CW–Eneolithic–like ancestry.
In addition to autosomal genetic changes through time, we observe a sharp reduction in Y-chromosomal diversity going from five different lineages in early CW to a dominant (single) lineage in late CW (Fig. 4A). We used forward simulations to explore the demographic scenarios that could account for the observed reduction in Y-chromosomal diversity. Performing 1 million simulations of a population with a starting frequency of R1a-M417(xZ645) centered around the observed starting frequency in Bohemia_CW_Early (3 of 11, 0.27), we assessed the plausibility of this lineage reaching the observed frequency in Bohemia_CW_Late (10 of 11, 0.91) in the time frame of 500 years under a model of a closed population and random mating (Materials and Methods). We reject the “neutral” hypothesis, i.e., that this change in frequency occurred by chance, given a wide range of plausible population sizes. Instead, our results suggest that R1a-M417(xZ645) was subject to a nonrandom increase in frequency, resulting in these males having 15.79% (4.12 to 44.42%) more surviving offspring per generation relative to males of other Y-haplogroups. We also find that this change in Y chromosome frequency is extreme compared to the changes in allele frequencies at fully covered autosomal 1240k sites (P < 0.0003) within the same males, suggesting a process that disproportionately affected Y-chromosomal compared to autosomal genetic diversity, ruling out a population bottleneck as the likely cause. Our results suggest that the Y-lineage diversity in early CW males was supplanted by a nonrandom process [selection, social structure, or influx of nonlocal R1a-M417(xZ645) lineages] that drove the collapse in Y-chromosomal diversity. A simultaneous decline of Y-chromosomal diversity dating to the Neolithic has been observed across most extant Y-haplogroups (64), possibly due to increased conflict between male-mediated patrilines (65). We view that changes in social structure (e.g., an isolated mating network with strictly exclusive social norms) could be an alternative cause but would be difficult to distinguish in the underlying model parameters.
|
|
|
|
Post by Admin on Aug 30, 2021 4:45:14 GMT
 Fig. 4 Temporal Y chromosome and autosomal PC2 variation in Bohemia. (A) Temporal distribution of Y chromosome haplogroups by culture. Schematic of phylogenetic relationships between Y chromosome lineages is shown along y axis. Dashed vertical lines demarcate respective (colored) cultural group into early and late phases. (B) Temporal variation in PC2 showing the genetic variation of males and females within each cultural group. The greatest genetic differentiation within early CW individuals can be found at Vliněves. The fst value between the three highest and three lowest early CW individuals on PC2 from Vliněves is greater than pairwise comparisons of all modern-day European populations (Fig. 5 and table S27).  Fig. 5 Genetic distances of Early CW individuals from Vliněves. Pairwise Fst between the three highest (VLI076, VLI088, and VLI090; VLI_High) and three lowest (VLI008, VLI079, and VLI009; VLI_Low) early CW on PC2 from Vliněves in the context of modern European pairwise Fst (2 SE plotted; table S27). Inset: PCA of 1141 modern West Eurasian individuals (gray points) on which early CW individuals from Vliněves (green symbols with black outline) were projected (see also Fig. 2B). The highly differentiated pairs of modern European populations labeled in the main figure are colored in the inset PCA. |
|
