|
Post by Admin on Nov 17, 2023 19:06:55 GMT
Results To explore the question of migration outside of cosmopolitan centres, we generated genome-wide data for 96 ancient individuals from eight sites (Table 1) in the Cambridgeshire region (Figure 1A) to an average coverage of up to 3.8× (median 0.037×). Mitochondrial haplogroup could be determined for 71 individuals and a subset of 52 genomes which had autosomal coverage >0.01× were used in autosomal analyses including imputation-based analyses relying on 43 genomes (Data S1A). The individuals come from six sites across the Late Iron Age / Romano-British period and two earlier comparative sites from the Neolithic and Bronze Age (Table 1). The majority of our Romano-British sites were in use between 200–400 CE (Table 1, Data S1B) and encompass farmsteads and a cemetery with a number of burials thought to be decapitations (Wiseman et al. 2021) (Supplementary Note 1). In general, the genomes represent an equal distribution of males and females, as determined genetically, and a range of juveniles and adults of all ages (Data S1A). The average endogenous human DNA content varies by site, but overall is 12.03% and genome-wide coverage is 0.13×. Estimated contamination rates from mitochondrial DNA (mtDNA) range between 0 - 3.78% (median 0.83%) and the average misincorporation of C > T in the first five base pairs (bp) is 8.39% (Data S1A). Figure 1. Geographical and chronological distribution of the dataset and population affinities. (A) Site map. (B) timeline. (C) PCA based on imputed genomes.The proportion of total variation explained by PC1 and PC2 is 0.0005 and 0.0004, respectively. Roman and Early Medieval Period genomes shown with x and + symbols, respectively, include those reported in this study as well as those from Martiniano et al. 2016 and Schiffels et al. 2016.
|
|
|
Post by Admin on Nov 19, 2023 2:34:00 GMT
Population structure of Iron Age / Roman Cambridge We studied the ancestry of the Roman period genomes from Cambridgeshire in the context of other contemporary material from Britain and modern genomes from Europe and the Middle East using Principal Component Analysis (PCA) (Data S2). We found that Roman genomes from Cambridgeshire all draw their genetic ancestry from Western Europe (Figure 1C) and that similarly to the majority of Roman genomes from Yorkshire they cluster more closely with modern Welsh than local East Anglian genomes (Figure 1C). Unlike the Yorkshire individuals, we do not detect outliers among the 25 Cambridgeshire Roman genomes with >0.1× coverage examined. All Roman period populations examined show homogeneity in their North/West European ancestry in relation to external reference populations in PCA analyses based on imputed data (Figure 1, Figure S1) or projections made from haploid genotype calls (Figures S4 and S5).
We tested whether the imputed Roman genomes have different affinities to ancient and modern European populations using f4 statistics. Consistent with the increased Neolithic ancestry observed in Iron Age genomes from England by Patterson et al. (2022), all seven Roman Period sites we tested showed consistently higher drift sharing with Sardinian Neolithic genomes than genomes from Copper and Bronze Age England (-5.3<Z<-2.4; Figure 2A). All sites show higher affinity to Late Iron Age England than to Imperial Roman genomes (Figure 2B). Unlike the Roman Period York cemetery that included a burial of a long distance migrant from present-day Syria or Jordan (Martiniano et al. 2016), we find no evidence of long-distance migration from the Mediterranean region among the 33 imputed genomes from Roman Cambridgeshire that we tested (Figure S2A-C). The Cambridgeshire genomes are also not differentiated by their affinity to Late Iron Age genomes from France, Scotland and England (Figure S2D-E).
|
|
|
Post by Admin on Nov 20, 2023 19:37:51 GMT
Figure 1. Geographical and chronological distribution of the dataset and population affinities. (A) Site map. (B) timeline. (C) PCA based on imputed genomes.The proportion of total variation explained by PC1 and PC2 is 0.0005 and 0.0004, respectively. Roman and Early Medieval Period genomes shown with x and + symbols, respectively, include those reported in this study as well as those from Martiniano et al. 2016 and Schiffels et al. 2016. Population structure of Iron Age / Roman Cambridge We studied the ancestry of the Roman period genomes from Cambridgeshire in the context of other contemporary material from Britain and modern genomes from Europe and the Middle East using Principal Component Analysis (PCA) (Data S2). We found that Roman genomes from Cambridgeshire all draw their genetic ancestry from Western Europe (Figure 1C) and that similarly to the majority of Roman genomes from Yorkshire they cluster more closely with modern Welsh than local East Anglian genomes (Figure 1C). Unlike the Yorkshire individuals, we do not detect outliers among the 25 Cambridgeshire Roman genomes with >0.1× coverage examined. All Roman period populations examined show homogeneity in their North/West European ancestry in relation to external reference populations in PCA analyses based on imputed data (Figure 1, Figure S1) or projections made from haploid genotype calls (Figures S4 and S5). We tested whether the imputed Roman genomes have different affinities to ancient and modern European populations using f4 statistics. Consistent with the increased Neolithic ancestry observed in Iron Age genomes from England by Patterson et al. (2022), all seven Roman Period sites we tested showed consistently higher drift sharing with Sardinian Neolithic genomes than genomes from Copper and Bronze Age England (-5.3<Z<-2.4; Figure 2A). All sites show higher affinity to Late Iron Age England than to Imperial Roman genomes (Figure 2B). Unlike the Roman Period York cemetery that included a burial of a long distance migrant from present-day Syria or Jordan (Martiniano et al. 2016), we find no evidence of long-distance migration from the Mediterranean region among the 33 imputed genomes from Roman Cambridgeshire that we tested (Figure S2A-C). The Cambridgeshire genomes are also not differentiated by their affinity to Late Iron Age genomes from France, Scotland and England (Figure S2D-E).
|
|
|
Post by Admin on Nov 25, 2023 20:11:26 GMT
Figure 2. Genetic affinities of Roman Period sites in England to ancient and modern populations of Europe. A-B: affinities to ancient genome groups of individuals from the Allen Ancient DNA Resource v54. C-D: affinities to groups of 200 individuals from the UK Biobank born in France, Netherlands, Denmark and Scotland. Each plot shows the estimated f4 value with an error range of 2 standard deviations. Respective f4 plots by individuals of the Roman sites are shown in Figures S2-3. Red - Early Medieval Period, blue - Roman Period. Similarly to Roman Period genomes from York (Martiniano et al. 2016) we find higher proximity of the Cambridgeshire Roman genomes to present-day Dutch than French genomes (Figure 2C, Data S3) and no higher affinity with Danish than Scottish genomes akin to the Early Medieval genomes (Figure 2D). Neither do we observe any notable individual deviations from the patterns observed at site level (Figure S3A-B). In sharp contrast to the single outlier, the long-distance migrant in York, we observe relatively little difference in the affinities of the Roman genomes to present-day groups from England, Kent and East England, with a third of the Roman Period individuals from Cambridgeshire (East England) showing, however, minor but significantly higher affinity to present-day Kent than average present-day genomes from East England (Figure S3C). We further examined IBD sharing patterns between imputed genomes of Roman individuals from Cambridgeshire in context of available Roman Period data from York, Late Iron Age France (Fischer et al. 2022) and Early Medieval West Europe (Gretzinger et al. 2022) as well as UK Biobank data for individuals born in the UK and elsewhere in Europe (Data S4, Figure 3). Unsurprisingly we find a relatively high level of IBD sharing among geographically close Roman sites in Cambridgeshire with an average probability of 25% of individuals from one site sharing an IBD segment longer than 4cM with individuals from another site, which is more than twice as high as sharing among present-day individuals from East or Southeast England. However, the mean rate of IBD sharing of Cambridgeshire Roman genomes with geographically more distant Roman site from York is lower (23%, p=0.42 by two-tailed t-test) than the local average in context of lower (16%, p=0.0014) diachronic IBD sharing between Cambridgeshire Roman and Early Medieval sites. Notably, IBD sharing among Early Medieval sites from across England is higher (32%, p=0.002) than sharing among Roman sites in Cambridgeshire alone, remaining high for the English Early Medieval sites across the Channel with Early Medieval sites from Lower Saxony and the Netherlands (28%). Compared to Roman sites, the Early Medieval sites from East England show (p=5×10-7) increase in IBD sharing with present-day Scandinavian and Dutch genomes from approximately 10% to 15%, which is consistent with the major increase in that period of continental northern European ancestry detected by Gretzinger et al. (2022) (Gretzinger et al. 2022). At the same time, IBD sharing with Late Iron Age France drops in Cambridgeshire from the mean of 15.5% in the Roman to 10% in the Early Medieval and 8% in present-day East England which is comparable to the level of sharing between modern French and English (6.27%) (Data S4).
|
|
|
Post by Admin on Nov 26, 2023 20:01:52 GMT
Figure 3. Probabilities of IBD sharing among populations. Heatmap of probabilities of individuals from population i to share at least one IBD segment >4 cM with individuals from population j. Present-day population data from the UK Biobank, ancient imputed genomes include Late Iron Age of France (Fischer et al. 2022), Roman Period data from Cambridgeshire (this study), from York (Martiniano et al. 2016), Early Medieval data (Gretzinger et al. 2022). We calculated runs of homozygosity (ROH) using HapROH (Ringbauer et al. 2021). Using a two-tailed Student’s T-test we find no difference in the average sum of ROH segments greater than 4cM or 8cM between the Roman era sites (Data S5). Nor, do we find a difference between the two newly generated Bronze Age (Over Barrows) samples and the Roman era populations (Data S5). We do find a difference between our Roman Period populations and the samples from York (4cM p = 0.014; 8cM p = 0.034), but only when comparing our imputed data (haploid mode) to the data available in the Allen Ancient DNA dataset (Anon). When comparing our imputed data to the York ROHs imputed in haploid and diploid mode from the shotgun data, there is no difference detected at either 4cM or 8cM lengths. Uniparental marker diversity To determine variation in the paternal lineages we called the genotypes of 161,140 Y chromosome haplogroup informative binary markers in 29 male samples from Late Iron Age and Roman Cambridgeshire with autosomal coverage >0.01× (Data S1C). All individuals could be assigned to haplogroups common in modern-day Europe. Majority of the samples (85%) belong to haplogroup R1b which became the predominant male lineage in Britain after the spread of the Beaker complex (Data S1D). Two first-degree related individuals from Duxford fall into the I2 clade which captures all previously known Y chromosome lineages in Britain before the Bell Beaker Culture (Data S1D-E). It is not clear, however, whether this particular lineage (I2-Y3722) of the Duxford father-son pair, reflects local continuity and survival from a pre-Beaker population as its present-day distribution is mainly focused on Ireland with only rare cases detected in England and Scotland (https://www.yfull.com/tree/I-Y3722/). Among R1b individuals with more coverage we identify distinct sub-clades, including the British/Irish Bell Beaker signature lineage R1b2-L21 (Patterson et al. 2021) as well as lineages from clades such as R1b11-Z2103 and R1b18-S1194 that have not been encountered in Britain in context of earlier time periods. Notably, none of the four R1b samples with >0.2× Y chromosome coverage fall to the same sub-clade. Some of the identified subclades of R1b appear to be rare in a large, high resolution modern Y chromosome compendium of more than 60,000 FamilyTreeDNA customers (Data S1D). Overall, compared to Copper/Bronze Age periods we do not detect in our Roman Cambridgeshire samples any notable changes in the composition of the Y chromosome haplogroups apart from a single I1 (NWC010) and a single G2a (DUX006) lineage that, by their presence in the Iron Age data by (Patterson et al. 2021), were likely introduced to Britain from the mainland in the Iron Age. (Data S1E). In autosomal DNA and isotope analyses these two individuals (NWC010, DUX006) did not stand out as outliers with external origins or high mobility during their lifetime (Figure 1C, Tables 2-3). We determined mitochondrial (mtDNA) haplotypes for 71 individuals (Data S1F) and found high diversity, with no haplotype matches except in the case of close kinship (Data S1G).
|
|