|
Post by Admin on Aug 29, 2023 19:05:19 GMT
Fig. 4: Genetic change in Britain in the context of Europe-wide trends. (A) Eight ancient DNA time transects for up to four periods, plotting the mean of the EEF inference on the y-axis and on the x-axis using the average of dates of individuals in periods defined for each region as in Supplementary Table 5. Sample sizes used to compute each point are given in Supplementary Table 7. Dotted lines connecting points should not be interpreted as implying a smooth change over time and instead are meant to help in visual discernment of which groups of points come from the same time transects. (B) The allele conferring lactase persistence experienced its major rise about a frequency millennium earlier in Britain than in Central Europe suggesting different selection regimes and possibly cultural differences in the use of dairy products in the two regions in the IA. This analysis based on imputed data includes 459 ancient individuals from Britain and 468 from Central Europe (Czech Republic, Slovakia, Croatia, Hungary, Austria, Germany and Slovenia) (we then co-analyzed with present-day individuals; Methods). Each vertical bar represents the derived allele frequency for each individual with values [0, 0.5, 1]; we use jitter on the x-axis, and show in shading the inferred 95% confidence interval for the allele frequency at each time point. Demographic change in Britain is also evident from another aspect of the data: the rate of runs of homozygosity (ROH), which can occur when a person’s parents are closely related. The larger the pool of people from which individuals draw their mates, the less likely it is for parents to be closely related, and thus we can average the number of 4–8 centimorgan (cM) ROH segments to estimate the effective size of the pool of people within which people were mating in the ~600 year period prior to the time when the analysed individuals lived17. We find that the size of the mating pool increased by roughly four-fold from the Neolithic to the IA (Extended Data Fig. 3), but this should not be interpreted as an estimate of census population size changes over this period as mating pool sizes are also affected by changing social customs. First, if the distance over which people ranged to find their mates was higher in some cultural contexts than in others, it would cause mating pool sizes to be different even if there was no difference in population densities; for example, mating pool size may have been less than the island-wide population size if members of communities mixed little with their neighbours16, or larger if individuals mated not only with people outside their local communities but also outside Britain. Second, we have gaps in sampling, especially at the end of the Neolithic (roughly 3000–2450 BCE), which means that demographic processes in such periods may be obscured. Third, due to the method effectively averaging mating pool size over centuries, this analysis may also fail to detect population declines over the space of a few decades.
|
|
|
Post by Admin on Aug 31, 2023 18:04:43 GMT
British change in European context We co-analysed our ancient DNA time transect in Britain alongside European transects (Fig. 4A, Supplementary Tables 5 and 7). Average EEF ancestry increased in North-Central Europe (Czech Republic/Slovakia/Germany) just as in Britain, with the first individuals with greatly increased EEF ancestry associated with artefacts traditionally classified as part of the Knoviz culture, a component of the broader Urnfield cultural complex (1300–800 BCE) that spread across much of Central Europe. This is particularly striking as the Knoviz individuals are from a population that is genetically similar to the Margetts Pit and Cliffs End Farm outliers (Supplementary Information section 6). Later individuals in North-Central Europe have similar EEF proportions, consistent with substantial continuity through the LBA-IA. In MBA and LBA France/Switzerland and South-Central Europe (Austria/Hungary/Slovenia) there was little change in average EEF ancestry, while EEF ancestry decreased in MBA and LBA Iberia (Spain/Portugal). There are two exceptions to this broad pattern of ancestry convergence in Europe—Scotland in the far north, and Sardinia in the far south—both of which have extreme and relatively unchanging proportions of EEF ancestry in this period (Supplementary Table 7).
This study multiplies by almost eight-fold the number of IA individuals with genome-wide data from Western and Central Europe (from 80 to 624; Supplementary Table 5), making it possible to accurately track the frequency change of genetic variants into the IA (Supplementary Table 8). Variants associated with light skin pigmentation at SLC45A2 became substantially more common throughout Europe in the IA. We obtain an unexpected result for the derived allele at MCM6-LCT rs4988235 which is associated with lactase persistence into adulthood (Extended Data Fig. 4). Previous analyses found that its frequency in the IA in sampled parts of continental Europe was a small fraction of its present-day incidence18. We document this at high precision in our dataset in Iberia where it was ~9% compared to ~40% today, and in Central Europe (Austria, Hungary, Slovenia, Czech Republic, Slovakia and Germany) where it was ~7% compared to ~48% today. However, in IA Britain its frequency was 50% compared to the current 73%, showing that intense selection to increase the frequency of this allele acted roughly a millennium earlier in Britain than it did in multiple parts of continental Europe (Fig. 4B, Extended Data Fig. 4). We find no evidence that the frequency rise in Britain was due to M-LBA migration: the Margetts Pit and Cliffs End Farm outliers did not carry the allele, and most of the rise in Britain occurred after the M-LBA (Fig. 4B, Supplementary Table 8). This suggests that dairy products were consumed in a qualitatively different way or were economically more important in LBA-IA Britain than in Central Europe.
Continental sources of M-LBA migration The ancestry change in Britain during the M-LBA was more subtle than those associated with the Neolithic and Beaker-period migrations. In England and Wales, allele frequency differentiation between the Neolithic and C/EBA was FST~0.02, but between the C/EBA and the IA it was an order of magnitude smaller at FST~0.002 (Extended Data Table 1). The pre-LBA population in Britain also made a substantial genetic contribution to the IA population, in contrast to the two earlier major Holocene ancestry shifts8,9. Evidence for a substantial contribution from the C/EBA population to later populations also comes from Y chromosome haplogroup R1b-P312/L21/M529 (R1b1a1a2a1a2c1), which is present at 89±5% in sampled individuals from C/EBA Britain and is nearly absent in available ancient DNA data from C/EBA Europe (Supplementary Table 9). The haplogroup remained more common in Britain than in continental Europe in every later period, and continues to be a distinctive feature of the British isles as its frequency in Britain and Ireland today (14–71% depending on region19) is far higher than anywhere else in continental Europe (Extended Data Fig. 5).
To gain insight into the possible sources of the M-LBA migrants to southern Britain, we fit the pooled IA individuals from England and Wales in qpAdm as a mixture of the main C/EBA cluster, and a second source. We tested 65 second sources—63 from continental Europe and 2 from Britain (the Margetts Pit outlier pool, and the Cliffs End Farm outlier pool)—and found that 20 fit at p>0.05. We then pooled the genetically similar Margetts Pit and Cliffs End Farm individuals and performed further testing with more stringent qpAdm setups, leaving eight second sources that consistently fit well with modest standard errors (Table 2, Supplementary Information section 6). The Margetts Pit and Cliffs End Farm pool fit as contributing 49.4±3.0% of the ancestry of IA people from southern Britain. Even omitting representatives of the putative source population living in Britain itself, we infer large genetic turnovers, as the seven continental populations that fit as sources are estimated to contribute 24–69% ancestry. Although only 1/5th of the continental candidate populations we tested are from France, 6/7th’s of the fitting populations are: four from Occitanie in southern France (600–200 BCE), two from Grand Est in northeastern France (800–200 BCE), and one from Spain (a ~600 BCE group). These fitting second sources all significantly post-date the ancestry change in Britain and hence cannot be the true sources; however, they are plausibly descended from earlier local populations. An origin in France is also suggested by the fact that all of the high EEF outliers in Britain in the M-LBA, and all of the 1000–875 BCE individuals that track the ramp-up of EEF ancestry from MBA to IA levels, are from Kent in far southeastern Britain (Extended Data Fig. 6). The migrant stream began admixing more broadly through southern Britain by the second half of the LBA, as individual I12624 from Blackberry Field, Potterne in Wiltshire, dated to 950–750 BCE, had an EEF proportion of 38.1±2.0% consistent with the level that became ubiquitous in southern Britain by the beginning of the IA (Extended Data Fig. 3). However, as this is the only non-Kent datapoint from the second half of the LBA, more sampling is needed to understand the geographic and temporal course of the spread of this ancestry.
|
|
|
Post by Admin on Sept 1, 2023 18:37:08 GMT
Continental sources of M-LBA migration The ancestry change in Britain during the M-LBA was more subtle than those associated with the Neolithic and Beaker-period migrations. In England and Wales, allele frequency differentiation between the Neolithic and C/EBA was FST~0.02, but between the C/EBA and the IA it was an order of magnitude smaller at FST~0.002 (Extended Data Table 1). The pre-LBA population in Britain also made a substantial genetic contribution to the IA population, in contrast to the two earlier major Holocene ancestry shifts8,9. Evidence for a substantial contribution from the C/EBA population to later populations also comes from Y chromosome haplogroup R1b-P312/L21/M529 (R1b1a1a2a1a2c1), which is present at 89±5% in sampled individuals from C/EBA Britain and is nearly absent in available ancient DNA data from C/EBA Europe (Supplementary Table 9). The haplogroup remained more common in Britain than in continental Europe in every later period, and continues to be a distinctive feature of the British isles as its frequency in Britain and Ireland today (14–71% depending on region19) is far higher than anywhere else in continental Europe (Extended Data Fig. 5).
To gain insight into the possible sources of the M-LBA migrants to southern Britain, we fit the pooled IA individuals from England and Wales in qpAdm as a mixture of the main C/EBA cluster, and a second source. We tested 65 second sources—63 from continental Europe and 2 from Britain (the Margetts Pit outlier pool, and the Cliffs End Farm outlier pool)—and found that 20 fit at p>0.05. We then pooled the genetically similar Margetts Pit and Cliffs End Farm individuals and performed further testing with more stringent qpAdm setups, leaving eight second sources that consistently fit well with modest standard errors (Table 2, Supplementary Information section 6). The Margetts Pit and Cliffs End Farm pool fit as contributing 49.4±3.0% of the ancestry of IA people from southern Britain. Even omitting representatives of the putative source population living in Britain itself, we infer large genetic turnovers, as the seven continental populations that fit as sources are estimated to contribute 24–69% ancestry. Although only 1/5th of the continental candidate populations we tested are from France, 6/7th’s of the fitting populations are: four from Occitanie in southern France (600–200 BCE), two from Grand Est in northeastern France (800–200 BCE), and one from Spain (a ~600 BCE group). These fitting second sources all significantly post-date the ancestry change in Britain and hence cannot be the true sources; however, they are plausibly descended from earlier local populations. An origin in France is also suggested by the fact that all of the high EEF outliers in Britain in the M-LBA, and all of the 1000–875 BCE individuals that track the ramp-up of EEF ancestry from MBA to IA levels, are from Kent in far southeastern Britain (Extended Data Fig. 6). The migrant stream began admixing more broadly through southern Britain by the second half of the LBA, as individual I12624 from Blackberry Field, Potterne in Wiltshire, dated to 950–750 BCE, had an EEF proportion of 38.1±2.0% consistent with the level that became ubiquitous in southern Britain by the beginning of the IA (Extended Data Fig. 3). However, as this is the only non-Kent datapoint from the second half of the LBA, more sampling is needed to understand the geographic and temporal course of the spread of this ancestry.
Table 2: Fitting proxies for the new ancestry source in Iron Age southern Britain
Proxies for source of the new ancestry N Mean date p-value Ancestry Margetts Pit and Cliffs End Farm M-LBA 4 1036 BCE 0.07 49.4 ± 3.0% Spain IA Tartessian 2 629 BCE 0.16 23.7 ± 1.2% France GrandEst IA1 (shotgun data) 5 620 BCE 1.00 48.9 ± 3.7% France Occitanie IA2 (high EEF subgroup, shotgun data) 1 450 BCE 0.85 25.8 ± 1.7% France Occitanie IA2 (high WHG subgroup, shotgun data) 1 450 BCE 0.39 33.5 ± 4.1% France Occitanie IA2 (shotgun data) 2 400 BCE 0.25 53.3 ± 5.4% France Occitanie IA2 (low Steppe subgroup, shotgun data) 2 363 BCE 0.33 36.5 ± 2.6% France GrandEst IA2 12 250 BCE 0.09 68.5 ± 3.3%
Note: We fit the pooled IA individuals from England and Wales as a mixture of the pooled C/EBA individuals from England and Wales and a proxy for the new ancestry source. The p-value is from qpAdm’s test of fit of each population as a two-way admixture with no correction for multiple hypothesis testing. These results represent eight of the 65 lines in Supplementary Information section 6, Table S6.1
|
|
|
Post by Admin on Sept 2, 2023 19:09:01 GMT
Regional variation in Iron Age Britain Estimates of Margetts Pit and Cliffs End Farm-like ancestry in southern Britain range from 35±5% in northern England to 56±5% in south-central England (Table 1, Extended Data Table 2). The IA was a period when material culture was increasingly regional in character16, and our results show that this was accompanied by subtle genetic structure, although without southern Britain there is no clear correlation of these admixture proportions to latitude (Table 1). We highlight the case of East Yorkshire, where most individuals are from ‘Arras Culture’ contexts comprising square-ditched barrows and occasional chariot burials. Similarities to funerary traditions of IA societies in the Paris Basin and Ardennes/Champagne regions have led to suggestions that East Yorkshire was influenced by direct migration from continental Europe in the IA20. Our estimate of the Margetts Pit/Cliffs End Farm ancestry source for East Yorkshire burials is 44±4% (Table 1), typical for middle latitudes of Britain at this time (East Anglia is similar). However, the East Yorkshire burials are distinctive in another way: regional differentiation in IA Britain, as measured by FST, is higher between East Yorkshire and other groups than it is between any other pair of IA populations in England and Wales in our dataset (Extended Data Table 2). Comparative data from the continent could make it possible to determine if this is due to isolation of IA East Yorkshire from the rest of southern Britain, or later streams of migration specifically affecting East Yorkshire.
Archaeological and linguistic context The period from 1500–1150 BCE has long been recognized as a time when cultural connections between Britain and regions of continental Europe intensified, and when societies on both sides of the Channel shared cultural features including domestic pottery, metalwork and ritual depositional practices2–6. From around 750 BCE there is more limited archaeological evidence of contact between Britain and the continent, and our genetic findings concur in showing that, by the beginning of the IA, there is little evidence of demographically significant migration into Britain2. Our findings do not establish whether the population movements we infer were a cause or consequence of M-LBA exchange networks, but they do suggest that interactions between local populations of Britain and new migrants bringing ideas from continental Europe could have been a vector for some of the cultural change we see in M-LBA England and Wales. Western and Central France are much more poorly represented by available genome-wide ancient DNA data than neighboring regions of Europe, and thus we cannot at present test if the gene flow between the two regions in this period was largely unidirectional.
Population movements are often a significant driver of cultural change, including in the languages people speak. While periods of intense migration such as the one we infer here do not always result in language shifts18, genetic evidence of significant migration is important because it documents demographic processes that are plausible conduits for language spread21. Several researchers have interpreted linguistic data as providing evidence for early Celtic languages spreading into Britain from France at the end of the Bronze Age or in the early IA22,23. Our identification of substantial migration into Britain from sources that best fit populations in France provides an independent line of evidence in support of this, and points to the M-LBA as a prime candidate for the period of this language spread. While the lack of evidence for M-LBA EEF ancestry change in Scotland could be interpreted as weakening the case that Celtic language spread into Britain at this time, a later arrival of Celtic languages in Scotland is consistent with evidence that non-Celtic and Celtic languages coexisted there into the first millennium CE24. Our finding of a decrease of EEF ancestry in Iberia, where the proportion was relatively high in the EBA, and a roughly simultaneous increase in Britain where the proportion was relatively low in the EBA (Fig. 4a), could, in theory, reflect a Celtic-speaking group of people with intermediate EEF ancestry spreading into both regions, although such a simple model cannot explain all the north-south ancestry convergence in Europe (Supplementary Information section 7). Nevertheless, the fact that the Margetts Pit and Cliffs End Farm outliers are genetically very similar to the Knoviz culture sample from Central Europe (Supplementary Information section 6) is striking in light of the fact that some scholars have hypothesized Central European Urnfield groups like Knoviz to have links to Celtic language spread25. Our failure to find evidence of large-scale migration into Britain from continental Europe in the IA suggests that, if Celtic language spread was driven by large-scale movement of people, it is unlikely to have occurred at this time. The adoption in IA Britain of cultural practices originating in continental Europe—particularly those linked to the La Tène tradition26—was also evidently independent of large-scale population movements, although there certainly were smaller movements, attested by individual IA outliers with high EEF ancestry such as those at Thame or Winnall Down (Fig. 3).
An important direction for future work is to generate new ancient DNA data from continental contexts especially in central and western France—and also Ireland—to test the alternative scenarios of population history consistent with the observations in this study, and to develop theories integrating the genetic findings within archaeological frameworks.
|
|