|
Post by Admin on Sept 28, 2022 7:42:17 GMT
These genetic patterns suggest that early medieval individuals from England have variable amounts of CNE ancestry. Although most individuals from early medieval English sites cluster clearly with either present-day WBI samples or CNEs, many individuals fall between these two clusters, suggesting admixture between these ancestral groups. To quantitatively estimate these ancestry compositions, we decomposed ancestral sources using a supervised clustering approach implemented in the software ADMIXTURE48. Specifically, we assembled modern populations into two metapopulations that serve as proxies for the source ancestries in early medieval England defined above: CNE (n = 407) and WBI (n = 667). We confirmed that these two present-day metapopulations accurately represent the ancient admixture sources by testing their relationships to the ancient individuals from England using FST statistics and F4 statistics of the form F4(Yoruba, Test; WBI, CNE) (Extended Data Fig. 1 and 2). The resulting ancestry estimates for early medieval English individuals are indeed tightly congruent with both PCA PC1 position and F4 statistics (Pearson's |r| > 0.9 between PCA, F4 and ADMIXTURE ancestry assessments). Applying our CNE–WBI ancestry decomposition to prehistoric samples, we found the genome-wide CNE ancestry in Britain and Ireland to be very low before the Early Middle Ages (Extended Data Fig. 3). In Bell Beaker and Bronze Age individuals from England, CNE ancestry does not account for more than 1% (Fig. 3a). This cannot be explained by genetic drift due to the temporal gap between our present-day CNE proxy and the Bronze Age, as shown by F4 statistics (Extended Data Figs. 1 and 2), which are robust against such drift. Similar proportions were also measured during the Iron Age (1% on average). CNE ancestry increased only during the Roman period, to 15% on average, although this estimate is based on only seven individuals. Six of those seven Roman-era individuals are from a single site, Eboracum (present-day York); which was a Colonia, the highest rank of Roman city with a legionary fortress, and as such it may have attracted a more cosmopolitan population than most of the rest of England (Fig. 3b). Fig. 3: Individual-based and site-based ancestry decomposition.
|
|
|
Post by Admin on Sept 30, 2022 5:52:46 GMT
In contrast to these previous periods, the majority of the early medieval individuals from England in our sample derive either all or a large fraction of their ancestry from continental northern Europe, with CNE ancestry of 76 ± 2% on average (Methods). Although CNE ancestry is predominant in central and eastern England, it is much less prevalent in the south and southwest of England, and absent in the one site that we analysed from Ireland (Fig. 3b). Moreover, we observed differences in continental ancestry not only between but also within sites. Although we estimate CNE ancestry to be prevalent across eastern English cemeteries, there was considerable variation at the individual level, ranging from 0% to 100% of CNE ancestry within a site. For example, at Hatherdene Close (n = 17) in Cambridgeshire, we estimated a mean CNE ancestry of approximately 70%, with eight individuals exhibiting exclusively CNE ancestry, but three individuals having low or zero CNE ancestry. Overall, these patterns of genetic heterogeneity, from the transregional to the family level, are consistent with continuous interaction between the Iron Age-derived Romano-British population and migrants from the continent.
We find no significant differences of CNE or WBI ancestry between male and female individuals (Supplementary Note 7), and find individuals of both ancestries within prominent and/or furnished burials. In England overall, individuals with CNE ancestry (here and in the following, CNE means more than 50% CNE, and WBI means less than 50% CNE) are more likely to be found with grave goods than individuals with WBI ancestry (Fisher’s exact test P = 0.016). Of note, this appears to be driven by female individuals with CNE ancestry who are more likely to be found with grave goods (P = 0.001), and in particular brooches (P = 0.012), than female individuals with WBI ancestry (both based on Fisher’s exact test). However, graves belonging to male individuals with CNE ancestry are just as likely to have grave goods (P = 0.57) or weapons (P = 1) as those with WBI ancestry (both based on Fisher’s exact test). This is underlined by specific examples, such as a near 100% WBI male burial in grave 37 at Updown Eastry found with a seax under a barrow marked by a ring ditch, indicating a prominent weapon burial associated with a prominent person or status (Supplementary Fig. 1.1).
This pattern is also visible in East Anglia specifically, where individuals with CNE ancestry more often have grave goods (P = 0.014). This is also significant when considering only female individuals (P = 0.025), but not when considering females with brooches, which display gender-related status (P = 0.197). At the site level, these patterns are partly significant at Hatherdene Close (P = 0.015, 0.036 and 0.1, respectively). Treating ancestry not as a binary but as a continuous variable largely agrees with the previous results (see Supplementary Note 7), with a notable exception of West Heslerton, which stands out from this overall pattern, where men with a greater proportion of CNE ancestry are more likely to be found with weapons (Wilcoxon rank sum P = 0.02, although non-significant with Fisher’s exact test P = 0.53), which is the only significant signal of this type that we found (Lakenheath also displays many CNE burials with weapons, but with limited sample size).
There are notable individual exceptions to these patterns, such as a predominantly (60%) WBI burial at grave 80 in Oakington, found with the skeleton of a cow, silvered disc brooches and a chatelaine, and interred under a barrow, which is one of the more notable or wealthy burials in this cemetery49 (Supplementary Fig. 1.4). We note that several burials with weapons that were previously identified as female and discussed in the literature23,50 have turned out to be genetically male in our analysis (see the highlighted entries in Supplementary Table 1). Of note, however, a single individual still displays a sex–gender difference: a teenage boy buried with an equal-arm brooch, beads and a knife (grave 122 in West Heslerton).
In Dover Buckland, one of the most comprehensively sampled cemeteries in our dataset, we observed the mixing of genetic and cultural identities at the family level. For example, we found a group of relatives, spanning at least three generations, who all exhibit unadmixed CNE ancestry (Extended Data Fig. 4a,c). Down the pedigree, we then see the integration of a female into this group, who herself had unadmixed WBI ancestry (grave 304), and two daughters (graves 290 and 426), consequently of mixed ancestry. WBI ancestry entered again one generation later, as visible in near 50:50 mixed-ancestry grandchildren (graves 414, 305 and 425). Grave goods, including brooches and weapons, are in fact found on both sides of this family tree, pre-mixing and post-mixing (for example, in the youngest and mixed generation, we found both weapons, beads and pin, and their mother with a brooch). Although the first mixed generation is buried in close proximity to each other, the grandchildren are elsewhere on the site, although placed together (Extended Data Fig. 4b).
A quite different pattern is observed at Apple Down, which is among the most western sites that we have analysed. Here graves can be classified into distinct burial configurations according to orientation, location and frequency of artefacts. We found that burials with CNE ancestry are more often buried in configuration A (located towards the middle of the site and with east–west burial orientation) than in configuration B (located more towards the edges and with north–south orientation)49 (Fisher’s exact test P = 0.048). This shows that there is a significant difference within the treatment of individuals according to their ancestry, a finding very similar to those at early medieval cemeteries in Hungary and Italy with respect to northern versus southern European ancestry51.
|
|
|
Post by Admin on Oct 1, 2022 5:54:45 GMT
Ancestry sources across the North Sea Our new continental medieval data from regions bordering the North Sea provide a unique opportunity to further investigate the potential source of the CNE-related ancestry increase that we have described above (Supplementary Note 3). To this end, we first selected individuals who, according to our CNE–WBI decomposition, are of unadmixed CNE ancestry (CNE of more than 95%; from here from as England EMA CNE). For each site in the continental dataset, we then tested whether its individuals were genetically similar to the England EMA CNE group (n = 109) in terms of allele frequencies. Among the continental medieval groups analysed, sites from both northern Germany and Denmark are indeed indistinguishable from England EMA CNE individuals (Fig. 4). Consistently, England EMA CNE and medieval individuals from Lower Saxony exhibit almost identical genetic affinities and ancestry components (Extended Data Fig. 5 and Supplementary Fig. 3.2), possess the highest level of genetic similarity (based on F2, F3, F4 and FST statistics) (Extended Data Fig. 5 and Supplementary Fig. 3.8) and are symmetrically related to most ancient and modern populations (Supplementary Table 3.12). Together, this suggests that they are likely derived from the same source population. Using the LOCATOR52 software, which uses machine learning to map individuals into geographical space based on their genetic profiles, we infer a region spanning the northern Netherlands to the southernmost tip of Sweden as a putative source for the England EMA CNE ancestors, with a large proportion of individuals being assigned to Lower Saxony (see Methods) (Fig. 4). This similarity adds to previous evidence from the material culture and burial practices, especially between the Elbe-Weser region and the early Anglo-Saxon cemeteries, from which the archaeological migration discourse initially arose53. However, we also note the strong genetic homogeneity among most analysed sites in the northern Netherlands, northern Germany and Denmark (Supplementary Note 4), implying that, during the Early Middle Ages, the continental North Sea and adjacent western Baltic Sea area was a genetic continuum spanning most of the western North European plain without major geographical substructure (Supplementary Fig. 4.1,4.4). This, together with genetic backflow from the British–Irish Isles into continental Europe (Supplementary Table 4.2 and Supplementary Fig. 4.2,4.4), reflects the inferred linguistic history54 and precludes further identification of specific microregions that contributed gene flow to Britain. We note that, although our screening of plausible medieval continental sites is broad, it could overemphasize later developments of the genetic structure due to the increased replacement of cremation burials by inhumations on the continent. It also has a specific caveat in Scandinavia, where our medieval reference populations are mostly from Viking-era burials, which have diverse and mixed ancestries that may not be representative of the earlier populations there42,44. Fig. 4: Identifying continental source regions for immigrant ancestry in early Medieval England.
|
|
|
Post by Admin on Oct 2, 2022 5:55:09 GMT
Already during the Early Middle Ages, several individuals from multiple sites exhibit modest degrees of excess affinity (5.4%) to present-day individuals from the Scandinavian peninsula (Supplementary Fig. 6.2a), indicating additional sources. Although close cultural contacts to the Scandinavian peninsula are attested in the archaeological record3, we did not find this genetic variation to be geographically stratified within early medieval England (Supplementary Fig. 6.2b). This Scandinavian Peninsula-related ancestry increases substantially (to 30.6%) only during the Viking period (Supplementary Note 6).
Having established the close relatedness between specific continental regions and the individuals from early medieval England, we modelled the latter more directly using ancient source populations with the method qpAdm35. Specifically, we pooled ancient individuals in England by site and modelled each group as being admixed between two sources: one represented by pooled Iron Age/Roman period individuals from England, and the other represented by pooled early medieval individuals from Lower Saxony (from here known as LowerSaxony EMA). The resulting admixture proportions obtained in this way for early medieval sites in England are strongly correlated with our mean estimates from supervised ADMIXTURE above, as well as site-wise F4 statistics and mean PCA position (Pearson's |r| > 0.9 between all four ancestry assessments) (Extended Data Fig. 6).
Using this model, we detected an average of 86 ± 2% ancestry from Lower Saxony across all early medieval sites in England, only slightly higher than the 76 ± 2% estimated using present-day source populations and supervised ADMIXTURE. At a regional scale, we observed more ancestry from Lower Saxony in eastern England than in the southwest, consistent with ancestry arriving from the east, either in one event or over a continuous time period. Our estimate of genome-wide ancestry is supported by independent evidence for population turnover from uniparental markers (Supplementary Fig. 2.7). Before the Middle Ages, post-Neolithic individuals from Britain and Ireland carried overwhelmingly the major Y chromosomal haplogroup R1b-P312, especially the sub-haplogroup R-L21 (refs. 39,41), which in present day shows a cline across the region, with highest frequencies in the west55,56. By contrast, the early medieval population of England exhibits a substantial fraction of continental-derived haplotypes belonging to haplogroups R1b-U106, R1a-M420, I2a1-L460 and I1-M253, which are commonly found in northern and central Europe (and are also common among ancient continental individuals including the ones that we report). In particular, Y chromosomal haplogroups I1-M253 and R1a-M420 were absent from our Bronze, Iron and Roman Age British and Irish individuals, but were identified in more than one-third of our individuals from early medieval England. Overall, haplogroups absent in Bronze and Iron Age England represent at least 73 ± 4% of the Y chromosomes in our early medieval English sample, mirroring the turnover estimates from autosomal data. Similarly, mitochondrial genomes show evidence of female lineage population turnover from regions bordering the North Sea (Supplementary Note 2 and Supplementary Fig. 2.4).
Estimates of continental ancestry on the X chromosomes (Supplementary Fig. 2.11), as well as estimates of source origin of Y chromosomal haplogroups (Supplementary Fig. 2.16) point to no significant difference between male-specific lineages and autosomal admixture estimates (Supplementary Note 2). Although neither mitochondrial, Y chromosomal or X chromosomal data can exclude subtle levels of sex bias during the admixture (Supplementary Note 2), they are also consistent with a model of no sex bias, suggesting that the migrants included both men and women who mixed at similar levels with the local population. We note that absence of sex bias during the early medieval CNE–WBI admixture does not exclude possibilities for sex bias in the later admixture processes that caused the dilution of CNE ancestry in present-day England observed below.
|
|
|
Post by Admin on Oct 3, 2022 7:15:13 GMT
Recent population shifts in England Although the most prominent signal of admixture in early medieval England is the rise in ancestry related to medieval and modern continental northern Europe, we found that several English sites include genomes that could not be explained as products of admixture between the two hypothesized ancestral gene pools—England IA or LowerSaxony EMA—using qpAdm57. Instead, these genomes have additional continental western and southern European ancestry (Supplementary Note 5). This ancestry is genetically very similar to Iron Age genomes from France5,6 (France IA) (Extended Data Fig. 7, Supplementary Table 5.1 and Supplementary Fig. 5.3,5.4). The majority of this French Iron Age-derived ancestry is found in early medieval southeastern England, namely, at the sites of Apple Down, Eastry, Dover Buckland and Rookery Hill, where it constitutes up to 51% of the ancestry identified (Extended Data Fig. 8a and Supplementary Table 5.4).
The appearance of France IA-related ancestry in early medieval England anticipates a pattern that we also clearly see in the present-day English population structure, in which we found that the same two-way CNE–WBI model that fits most ancient English fails for the modern population (Supplementary Fig. 5.8,5.11). Indeed, the missing component in the modern English population appears to be represented well by France IA (Supplementary Table 5.2 and Supplementary Fig. 5.2).
Using qpAdm (Methods), most present-day Scottish, Welsh and Irish genomes can be modelled as receiving most or all of their ancestry from the British Bronze or Iron Age reference groups, with little or no continental contribution. By contrast, for all present-day English samples the simple two-way admixture model (England LIA + England EMA CNE) fails. By extending our model to a three-way with added France IA as a third component, we now obtain fitting models (Supplementary Fig. 5.11,5.21). We estimate that the ancestry of the present-day English ranges between 25% and 47% England EMA CNE-like, 11% and 57% England LIA-like and 14% and 43% France IA-like. There are substantial genetic differences between English regions (Fig. 5a), with less ancient continental ancestry (England EMA CNE or France IA related) evident in southwestern and northwestern England as well as along the Welsh borders (Fig. 5c). By contrast, we saw peaks in CNE-like ancestry of up to 47% for southeastern, eastern and central England, especially Sussex, the East Midlands and East Anglia. We found substantial France IA ancestry only in England, but not in Wales, Scotland or Ireland, following an east-to-west cline in Britain (Pearson’s |r| > 0.86), accounting for as much as 43% of the ancestry in East Anglia (Fig. 5d). Very similar results were produced using LowerSaxony EMA as a source for CNE ancestry (Extended Data Fig. 8b). One potential caveat in this analysis is our relatively sparse Roman sample from England, where we particularly lack samples from the south, which might have pre-existing France IA-related ancestry. We, therefore, turned to one of our early medieval sites, the post-Roman cemetery of Worth Matravers at the southern coast of Dorset, whose individuals have nearly no CNE ancestry (less than 6% on average), and thus may serve as a more temporally close proxy for post-Roman Britain before the arrival of CNEs. When used as a source in our model, we found that the estimates of France IA-related ancestry in present-day England changed by less than 3% on average across the regions (Fig. 5b), suggesting that France IA-related ancestry entered England to a substantial amount after the Roman period. We note that a model involving southern or western European-like ancestry in England has been previously proposed36 on the basis of present-day samples, but we can now go further and delineate this third component more clearly against the CNE-like immigrant gene pool making up the majority of the early medieval individuals from England that we studied.
|
|