Figure1.(a) Principal component analysis of modern Europeans (grey) and projected ancient Europeans. The full list of ancient samples with sample IDs, the associated culture or time period followed, and two-letter International Organization for Standar- dization codes for country can be found in table 1 and electronic supplementary material,table S7.(b) Geographical range of BAC/CWC and admixture modelling with qpAdm of BAC/CWC individuals in a three-source model:Anatolian Neolithic (orange), European hunter–gatherers (blue), and Yamnaya steppe herders (green). Individuals marked with an asterisk lived after the BAC/CWC time period. Map source: Sjögren et al. [39], published CC-BY.

The broad continental pattern that appears confirms previous results (figure 1a) [3]: (i) a clear separation between Early and Middle Neolithic farmers and Mesolithic hunter-gatherers from all over Europe [1,30,32–34,42,43], (ii) substructure among the hunter–gatherers roughly corresponding to an east–west gradient [1,8,29,31,32], with an exception in Scandinavia [29], and (iii) most Late Neolithic and Bronze Age indivi-duals overlapping with modern-day central and northern Europeans due to ad-mixture with incoming groups related to the Yamnaya herders from the Pontic steppe [1,32].

In Scandinavia, we see a clear separation between individuals from the three diffe-rent archaeological complexes, the FBC, PWC,and BAC, with individuals associated with the farming FBC (including the Rössberga individuals reported here) grouping with other Middle Neolithic farmer groups (figure 1a). Neolithic foragers associated with the Pitted Ware Culture (including ajv54) were genetically similar to Mesolithic Scandinavian hunter–gatherers with some trend towards farming groups, likely due to admixture between the two groups [9,29,33]. The individuals associated with BAC (including the two individuals from Bergsgraven) show a clear association with indivi-duals from other CWC contexts elsewhere in Europe. Notably, the oll007 individual from the megalithic site at Öllsjö (southern Sweden), who was not directly associa-ted with any CWC-related artefacts but overlaps with CWC chronologically, clusters with individuals from CWC contexts, as do the two individuals with later dates (oll009 and oll010) from the same site.

Similar to the BAC individuals,the newly sequenced individuals from the present-day Karlova in Estonia and Obłaczkowo in Poland appear to have strong genetic affini-ties to other individuals from BAC and CWC contexts across the Baltic Sea region [1, 2, 7, 8, 23,31,32] (figure 1a). Some individuals from CWC contexts, including the two from Obłaczkowo, cluster closely with the potential source population of steppe-related ancestry, the Yamnaya herders.Notably,these individuals appear to be those with the earliest radiocarbon dates among all genetically investigated individuals from CWC contexts. Overall, for CWC-associated individuals, there is a clear trend of decreasing affinity to Yamnaya herders with time (figure 2).

Download figure

Open in new tab

Download powerPoint

Figure 2. Correlation between f4(Chimp, LBK, YAM, X), where X is a CWC or BAC individual, and the date (BCE) of each individual (table 1 and [1,2,7,8,23,31, 32, 37]). This statistic measures shared drift between CWC and Linear Pottery Culture (LBK) as opposed to YAM and should increase with the higher proportion of Neolithic farmer ancestry in CWC and BAC. The pattern is not driven by spurious affinities between single nucleotide polymorphism (SNP) capture or shotgun data (electronic supplementary material, figures S9 and S10). Ages on the x-axis correspond to the mid-point of the interval for the date of each sample as reported in their original publication (electronic supplementary material, tables S7 and S8). (Online version in colour.)

(c) Admixture modelling of Corded Ware Culture

To investigate the spread of CWC-related people into Scandinavia and obtain a clea-rer picture of the ancestry proportions found in individuals associated with the CWC and the BAC,we conducted a supervised modelling of their ancestry from three sour-ces: Anatolian farmers, western European hunter–gatherers, and Yamnaya steppe herders. We estimated ancestry under this three-source model per individual using qpAdm [1]. The geographical pattern of ancestry is shown in figure 1b.This approach confirms that most ancestry in all CWC individuals originates from a group related to Yamnaya steppe herders (green), while smaller contributions from Western hunter–gatherers (blue) and Neolithic farmers (orange) are also widespread (figure 1b).

Individuals from BAC contexts in (modern-day) Sweden also show this pattern, with a large genetic component tracing its ancestry to Yamnaya steppe herders and with small ancestry components related to Neolithic farmers and hunter–gatherers. The other individuals who were contemporary with BAC but had unclear cultural contexts and who were buried in the Ölljsö megalith constructed many hundred years earlier (oll007), or found as a stray find (Ölsund) [9],show the same genetic profile as indivi-duals from typical BAC contexts in other parts of Sweden. Although the individuals associated with the BAC (including the oll007, oll009, oll010, and Olsund) have a large proportion of steppe ancestry, it is relatively low compared with most other indi-viduals from CWC contexts. The female (kar1) from the Karlova CWC burial in Esto-nia shows similar patterns, also displaying slightly less steppe ancestry than other CWC-associated individuals from Estonian sites [2,7]. By contrast, the CWC indivi-duals from Obłaczkowo in Poland (poz44 and poz81) show an extremely high pro-portion of steppe ancestry (greater than 90%), which is different from the later CWC-associated individuals excavated in Pikutkowo (Poland) [23], but similar to some other CWC-associated individuals from Germany, Lithuania, and Latvia [2,8,31].

Interestingly, these individuals with a large fraction of steppe ancestry have typically been dated to more than 2600 BCE, making them among the earliest CWC indivi-duals genetically investigated.This observation, i.e.early CWC individuals resembled (genetically) Yamnaya-associated individuals, while later CWC groups show higher levels of European Neolithic farmer ancestry (Pearson's correlation coefficient: −0.51, p = 0.006) (figure 2), suggests an initial dispersal that occurred rapidly.

(d) Admixture modelling of Battle Axe Culture

The positioning in the principal component analysis (PCA) (figure 1a) and the admix-ture modelling results (figure 1b) suggest ancestry from three different main ancest-ral groups in BAC: Anatolian farmers, European hunter–gatherers, and Yamnaya steppe herders.Direct candidates for contributions to BAC are the Scandinavian Middle Neolithic PWC and FBC plus a third population carrying high proportions of Yamnaya-related ancestry (YAM),like CWC. We used an explicit model to investigate contri-butions from predefined ancestral groups (qpWave of ADMIXTOOLS) [1,44] to model the BAC genetic make-up. We found that the BAC-associated individuals can be modelled as a combination of genetic material from these three groups (p= 0.80). In ad-dition to the three-source model, two models only using two sources were con-sistent with the data: the best fitting two-source model was FBC+ YAM (p = 0.86), while YAM + PWC would also fit the data (although trending toward low p-values;p = 0.07). These observations suggest that - to our statistical resolution - a direct PWC contri-bution to BAC is not needed in a model, but actual PWC admixture might have been small or there may have been indirect PWC contributions through PWC first mixing with FBC [34] who later contributed ancestry to BAC. Notably, using only the CWC population as the single source for BAC was consistent with the data in all ca-ses (p >0.05,except when using CWC-associated individuals from Latvia, CWC_LV), showing that the BAC individuals in Sweden could not have emerged without migrations from other CWC groups.

To find out if BAC represents an admixed group that came into Scandinavia or if (at least some of) the admixture took place in Scandinavia, we investigated the relation-ship among the Scandinavian BAC group and various CWC groups from other geo-graphical areas by constructing admixture graphs (figure 3). The BAC groups fit as a sister group to the CWC-associated group from Estonia (CWC_EE,electronic sup-plementary material,figure S8) but not as a sister group to the CWC groups from Po- land (CWC_PL, figure 3) or Lithuania (CWC_LT, electronic supplementary material) (|Z|>3),indicating some differences in ancestry between these CWC groups and BAC (CWC from Latvia, CWC_LV, and the CWC group from Germany, CWC_DE, had to be excluded from this analysis due to the number of SNPs being too low, see elect-ronic supplementary material). Supervised admixture modelling suggests that BAC may be the CWC-related group with the lowest YAM-related ancestry and with more ancestry from European Neolithic groups (figure 1b). Consequently, models, where BAC is an admixed group between a CWC group and additional ancestry from a group carrying ancestry-related European Neolithic groups, were consistent with the data for all three tested CWC groups (|Z| < 2.3, figure 3; electronic supplementary material, figures S7 and S8).

Figure 3.Admixture graph modelling BAC as a mix of a CWC source and an admixed Middle Neolithic group (worst |Z| < 1.68). CWC individuals from modern-day Poland (CWC_PL) chosen here, other groups shown in electronic supplementary material, figures S7 and S8. WHG, western hunter-gatherers; EHG, eastern hunter-gatherers. The sub-Saharan African Mbuti were used as an outgroup for this analysis.

Two potential scenarios were consistent with the data to explain the emergence of BAC in Scandinavia: (i) direct migration of CWC groups from the eastern Baltic or (ii) a migration of CWC groups from the southern Baltic Sea region mixing with FBC groups in Scandinavia. Future archaeogenetic studies could fill geographical and chronological gaps by including more samples that should help to distinguish between the scenarios.

3. Discussion

The gene pool of modern Europeans has been shaped by a number of prehistoric events and migrations. The CWC horizon represents one of the major demographic processes in the northern areas of Europe as this is associated with the first occurrence of steppe-related ancestry. Our results have implications for our view on the demographic development associated with the CWC in general, and the Scandinavian variety of the BAC specifically.

People buried in CWC contexts display a genetic ancestry component that was not present in northern and central Europe prior to the third millennium BCE.This ances-try component, often called ‘steppe ancestry’, probably traces back to the ‘Yamnaya expansion’ of herders associated with the Yamnaya Culture that dispersed into Eas-tern Europe from the Pontic-Caspian steppe around 3000 BCE [1,2,32]. This compo-nent makes up the largest proportion of the genetic ancestry in all sequenced BAC / CWC individuals around the Baltic Sea: from the modern-day countries Estonia, Lat-via, and Lithuania in the east; Poland and Germany in the southwest; and Denmark and Sweden in the northwest (figure 1b). One important observation is that the ear-liest CWC individuals analysed to date had the highest proportion of steppe ancestry (greater than 90%), while this proportion decreased in later individuals (figure 2).

This suggests a gradual process of admixture between incoming groups and local groups, such as the FBC groups in northern Europe that traced most of their genetic ancestry to Anatolian Neolithic farmers. This process was driven by in-coming males mixing mainly with local females [5,7,45,46].The admixture process is evident across the en-tire distribution of the CWC, even in regions such as the eastern Baltic coast where no FBC groups or genetically related groups have been found. Consequently, farmer-related ancestry must have arrived in the eastern Baltic region via migrations more recent than 2500 BCE [7–9].Possibilities include a gene-rally large exchange network across the entire CWC horizon or specific migrations into the eastern Baltic region. Potential source regions for the latter could be modern-day Poland or Sweden, where FBC groups predating the arrival of the CWC are found.

The paternal lineages found in the BAC/CWC individuals remain enigmatic. The ma-jority of individuals from CWC contexts that have been genetically investigated this far for the Y-chromosome belong to Y-haplogroup R1a, while the majority of sequen-ced individuals of the presumed source population of Yamnaya steppe herders be-long to R1b [1,2]. R1a has been found in Mesolithic and Neolithic Ukraine [1,31,32]. This opens the possibility that the Yamnaya and CWC complexes may have been structured in terms of paternal lineages—possibly due to patrilineal inheritance sys-tems in the societies [47,48]-and that genetic studies have not yet targeted the direct sources of the expansions into central and northern Europe.

The Scandinavian Middle Neolithic megalithic tombs are associated with the FBC. However, their reuse, indicated by artefacts common to the BAC and later periods, has been noted [17].The oll007 individual,buried in the FBC-associated Öllsjö mega- lithic tomb, but radiocarbon dated to the time period of the BAC, is genetically very similar to individuals from BAC contexts (e.g. Bergsgraven and Viby). Thus, although archaeologically the reuse of megalithic tombs was assumed earlier [17], our study may be the first direct link (using genetics) showing that indeed FBC-associated me-galithic tombs were used as burial places also for the people of the BAC. This could possibly also extend to the Danish Single Grave Culture (SGC) [49], as RISE61 [2], a male buried in the Kyndeløse passage grave and with a radiocarbon date overlap-ping with the BAC/CWC/SGC time period, also displays some steppe ancestry.

The BAC replaces the FBC in the southern parts of Scandinavia and was previously assumed to have been a cultural adaptation of existing groups [12,17]. We show in multiple individuals from different parts of Scandinavia that these groups (BAC) were part of the general CWC horizon, i.e. they too are the result of admixture of different groups tracing parts of their ancestry to European hunter–gatherers, Anatolian Neoli-thic farmers, and Yamnaya steppe herders. This implies that BAC groups were not the direct descendants of any of the groups that lived in the area previously or even contemporaneously—i.e. the groups associated with FBC or the PWC. We also note that the BAC group does not have a particular genetic connection with other eastern Baltic groups such as the Combed Ceramic Culture. The mixed ancestry of indivi-duals in BAC contexts is evident across all autosomal analyses, as well as in mito-chondrial haplogroups,but the paternal haplogroups stand out to some extent, show-ing a deviant,more extreme,pattern. This Y-chromosome pattern is, however, consis- tent with a male sex-biased migration and admixture process among the Yamnaya, and later CWC, groups [7,39,45,50].

As the individuals of the BAC complex cannot be modelled as direct genetic descen-dants of FBC or PWC groups, a migration into Scandinavia of people with a large proportion of steppe ancestry must have taken place.We were not able to find unam-biguous evidence for a specific source population by testing all other individuals as-sociated with the CWC that have been genomically investigated as potential ances-tors.The Scandinavian BAC group has more Neolithic farmer ancestry than pre-2600 BCE individuals in CWC contexts from the southern or eastern Baltic coast, sugges-ting that they mixed with an FBC group. This process could have happened in Scan-dinavia or before arriving in Scandinavia. On the one hand, the CWC people were mainly bound to travels on land, which would favour migration from central Europe into modern-day Denmark and Sweden (e.g. [16]). On the other hand, there is evi-dence for technological exchange (pottery craft) crossing the Baltic Sea in this time period [20,51],which may or may not be associated with gene flow. Finally, the gene-tic data from BAC contexts are still limited, and the patterns of gene flow that we ob-serve are consistent with both a single migration event into Scandinavia and with a continuous process with an extensive network of social and technological exchange. Future studies will refine our understanding of the social, geographical, and temporal dynamics during this important period in European prehistory.

4. Material and methods

We extracted DNA, prepared Illumina DNA libraries [29,33,34], and shotgun sequen-ced DNA from 11 archaeological remains from Sweden, Estonia,and Poland (table 1; electronic supplementary material). The samples were radiocarbon dated (n = 10) and analysed for stable isotope values (carbon,n = 8;nitrogen, n = 8; strontium, n = 2). Sequence data were merged, trimmed, and mapped to the human genome, and the data were authenticated using fragmentation, deamination patterns, and various con-tamination estimates [29]. Mitochondrial and Y-chromosomal DNA haplogroups were called for the newly sequenced individuals. The comparative data consisted of an-cient individuals with at least 0.1×genome coverage who lived at the time of the CWC or before (electronic supplementary material, table S7) [1,2,7–9,23, 29 – 34, 42, 43, 52 – 59] and modern populations from the Human Origins dataset [30] and the public data of the Simons Genome Diversity Panel [60]. Ancestry and admixture were assessed using principal component analyses, unsupervised ADMIXTURE [41], and admixture proportions were estimated using ADMIXTOOLS' qpWave and qpAdm [1, 44], and phenotypic variation was investigated [61,62]. More detailed descriptions of the archaeological samples and sites and the methods used can be found in the electronic supplementary material.

Data accessibility

The genome data for the 11 investigated individuals are available at the European Nucleotide Archive (accession number PRJEB32786).

Authors' contributions

H.M., T.G., A.G., J.S., and M.J. designed the research; H.M.,T.G., E.M.S., A.J., M.F., A.R.M., A.G., J.S., and M.J. performed the research; H.M., A.J., M.F., Ł.P., M.T., J.L., and J.S. contributed samples and conducted archaeological analyses; H.M., T.G., and E.M.S. analysed data; and H.M., T.G., A.G., J.S., and M.J. wrote the paper with input from all authors.

Competing interests

We declare we have no competing interests.

Funding

This work was supported by the Berit Wallenberg Foundation (grant no. BWS2011.0090 (M.F.)); the Swedish Research Council (grant nos 2017-02503 (H.M.), 2017-05267 (T.G.), 2013-1905 (M.J.,J.S.,and A.G.));Riksbankens Jubileums- fond (grant no. M13-0904:1 (M.J., J.S., and A.G.)) and Knut and Alice Wallenberg foundation (M.J., J.S., and A.G.). The Ministry of Science and Higher Education of the Republic of Poland funded the radiocarbon dates for poz44 and poz81 (grant no. N N109 287 137 (L.P.)).

Footnotes

^†These authors contributed equally to this study.

Electronic supplementary material is available online at https://doi.org/10.6084/m9.figshare.c.4682318.