|
Post by Admin on Nov 27, 2023 21:30:23 GMT
Figure 4. Relatedness between ancient Iron Age/Roman genomes by autosomal and X chromosome calculated in two ways. A) Mismatch probabilities. Each dot shown on the plot represents a pair of ancient genomes assessed (with READ) for their average pairwise differences normalised by population average. First, second and third degree boundaries for the autosomal relatedness are estimated as in Kuhn et al. 2018. The lower boundary for 99% autosomally unrelated pairs is shown on the x-axis for guidance. Dots with high transparency correspond to pairs with low aggregate SNP coverage. B). Proportion of pairs with autosomal 1-3rd degree relatedness per site. Notably within the relationships detected within the sites we find several triangular cases of relatedness with a female individual involved within more than one pair (e.g. Duxford DUX011 (female) related with DUX019 (male) and DUX001 (male)), or in case of North West Cambridge we find a relationship between three sampled male individuals (NWC004, NWC010, and NWC009) who appear to be related to each other through unsampled female(s) (either unexcavated or not sampled) because their pairwise X chromosomal differences are lower than population average despite the fact that they carry different mtDNA lineages (Figure 4). Genetically related individuals appear not to be clustered or buried next to each other: for example, the members of a Duxford family DUX011 (mother), DUX008 (father) and their son (DUX001) are all buried in different groups of burials (Figure S6) identified in the original site report (Lyons 2011). Or, similarly, in Vicar’s Farm related pairs of individuals were buried in different groups of burials (Evans and Lucas 2020) (pages 333-34 & 377). We further used IBIS to explore IBD sharing within and among Late Iron Age and Roman sites in Cambridgeshire. In case of all pairs of imputed individuals that were identified with READ as closely related we found multiple IBD segments supporting their close relatedness (Data S4). However, in all cases the observed total IBD shared expressed in kinship coefficient was less than expected from the 1st-3rd degree relationship suggesting that capturing long tracts of IBD at low coverage is hindered by their fragmentation due to imputation errors. Besides the kinship pairs already detected with READ (Figure 4) we did not find any new relationships with IBIS within the sites. However, we detected a case of distant relatedness between DUX019 from Duxford and a previously reported sample 12884A (HI2, Schiffels et al. 2016) from Hinxton, who share five IBD segments longer than 7cM consistent with estimated kinship coefficient suggesting 6th degree relatedness. Given the Duxford and Hinxton sites are located only 3 kilometres from each other and are both in the Cam valley this finding points to local mobility between geographically adjacent sites (Data S4). Phenotypic Changes / Health To investigate potential changes in Cambridgeshire through the Roman Period, we imputed 114 SNPs known to be involved in phenotypic traits related to diet (carbohydrate, lipid and vitamin metabolism), immunity (response to pathogens, autoimmune diseases and other immunity traits) and pigmentation (eye, hair and skin colour) in the ancient individuals presented here plus 234 individuals from the literature, for a total of 277 samples divided into four groups from the Mesolithic to the Roman Period (Data S6A-D). Within Great Britain, from the Neolithic to modern Great Britain (1000 Genomes GBR) there are 34 significant SNPs (Data S6B). We can see two main “break-points” considering the significant group pairs after the Tukey test for each of the 34 SNPs, in line with previous findings: 1) after Neolithic and 2) after the Bronze Age. Most of the significant SNPs involve Neolithic or Chalcolithic/Bronze Age groups that differ from later periods. More specifically, they are involved in several different diet, immunity and pigmentation functions. In the diet group, we found six significant SNPs: two that confer lactase persistence, one involved in lipid metabolism, two in fatty acid metabolism and one in vitamin D metabolism. The SNP involved in the lipid metabolism is in an introgressed Denisovan-tract and tends to increase over time, with greater shifts after Neolithic and Chalcolithic/Bronze Age. The fatty acid metabolism SNPs show a change in frequency mainly after Chalcolithic/Bronze Age, while the SNP that is a protective factor against vitamin D deficiency show the lowest frequency during Chalcolithic/Bronze Age (0.49 vs. 0.67 in Neolithic and 0.72 on average in previous and later periods respectively). When focusing on the group pairs involving Iron Age/Romans (IAR), we see eight SNPs are different between IAR and modern GBR. The MCM6 locus, with two Lactase-persistence SNPs, show a sharp increase after Iron Age/Romans, after the first great increase after Bronze Age; this is consistent with recent findings related to low frequency of LCT alleles in Bronze Age and increase in frequency in later periods (e.g. (Burger et al. 2020; Segurel et al. 2020)). Between different sites of Roman UK and Roman Italy, no significant SNPs appear. Differently from (Kerner et al. 2021), we do not observe frequency fluctuations for the TB risk factor rs34536443, which is low in frequency from the Neolithic with no significant changes over time. It reached present-day frequency from IAR. Mobility through isotopic analysis To further explore potential childhood origins and geographical mobility, we measured oxygen isotope ratios in the tooth enamel of Iron Age and Roman individuals excavated from the Cambridgeshire region. The oxygen isotope composition of local water sources is largely determined by the local climatic conditions (Dansgaard 1964; Pederzani and Britton 2019). The oxygen isotope ratios measured in archaeological human tooth enamel are a reflection of the water consumed during the formation of the enamel during childhood and can therefore provide information about the environment the individual grew up in (DeNiro and Epstein 1978; Longinelli 1984; Luz, Kolodny, and Horowitz 1984; Luz, Kolodny, and Kovach 1984). This can be broadly used to identify population mobility if this does not match the expected values for the environment in which they are buried (Pederzani and Britton 2019). We measured the carbonate oxygen isotope ratios (δ18OCO3) of 33 second premolars from 17 individuals (1 Early Iron Age, 1 Middle Iron Age, 15 Late Iron Age/Early Roman) from Duxford and 15 individuals (all Mid-Late Roman) from Vicar’s Farm, Cambridge. We compared the results to published δ18OCO3 data from 33 individuals from Knobb’s Farm (1 Middle Iron Age, 32 Late Roman: (Wiseman et al. 2021). The datasets are broadly comparable, although due to the variety of teeth analysed, the data will not represent exactly the same period of life (Lightfoot in Wiseman et al. 2021, p. 160). The δ18OCO3 values across the three sites are wide-ranging and overlapping (Figure 5) (Data S7). The distribution of the δ18O data slightly differs. However, the range of values seen at Duxford and Vicar’s Farm fit well within the distribution of the data from Knobb’s Farm (Figure 5).
|
|
|
Post by Admin on Nov 30, 2023 16:41:13 GMT
Phenotypic Changes / Health To investigate potential changes in Cambridgeshire through the Roman Period, we imputed 114 SNPs known to be involved in phenotypic traits related to diet (carbohydrate, lipid and vitamin metabolism), immunity (response to pathogens, autoimmune diseases and other immunity traits) and pigmentation (eye, hair and skin colour) in the ancient individuals presented here plus 234 individuals from the literature, for a total of 277 samples divided into four groups from the Mesolithic to the Roman Period (Data S6A-D). Within Great Britain, from the Neolithic to modern Great Britain (1000 Genomes GBR) there are 34 significant SNPs (Data S6B). We can see two main “break-points” considering the significant group pairs after the Tukey test for each of the 34 SNPs, in line with previous findings: 1) after Neolithic and 2) after the Bronze Age. Most of the significant SNPs involve Neolithic or Chalcolithic/Bronze Age groups that differ from later periods. More specifically, they are involved in several different diet, immunity and pigmentation functions. In the diet group, we found six significant SNPs: two that confer lactase persistence, one involved in lipid metabolism, two in fatty acid metabolism and one in vitamin D metabolism. The SNP involved in the lipid metabolism is in an introgressed Denisovan-tract and tends to increase over time, with greater shifts after Neolithic and Chalcolithic/Bronze Age. The fatty acid metabolism SNPs show a change in frequency mainly after Chalcolithic/Bronze Age, while the SNP that is a protective factor against vitamin D deficiency show the lowest frequency during Chalcolithic/Bronze Age (0.49 vs. 0.67 in Neolithic and 0.72 on average in previous and later periods respectively).
When focusing on the group pairs involving Iron Age/Romans (IAR), we see eight SNPs are different between IAR and modern GBR. The MCM6 locus, with two Lactase-persistence SNPs, show a sharp increase after Iron Age/Romans, after the first great increase after Bronze Age; this is consistent with recent findings related to low frequency of LCT alleles in Bronze Age and increase in frequency in later periods (e.g. (Burger et al. 2020; Segurel et al. 2020)). Between different sites of Roman UK and Roman Italy, no significant SNPs appear. Differently from (Kerner et al. 2021), we do not observe frequency fluctuations for the TB risk factor rs34536443, which is low in frequency from the Neolithic with no significant changes over time. It reached present-day frequency from IAR.
Mobility through isotopic analysis To further explore potential childhood origins and geographical mobility, we measured oxygen isotope ratios in the tooth enamel of Iron Age and Roman individuals excavated from the Cambridgeshire region. The oxygen isotope composition of local water sources is largely determined by the local climatic conditions (Dansgaard 1964; Pederzani and Britton 2019). The oxygen isotope ratios measured in archaeological human tooth enamel are a reflection of the water consumed during the formation of the enamel during childhood and can therefore provide information about the environment the individual grew up in (DeNiro and Epstein 1978; Longinelli 1984; Luz, Kolodny, and Horowitz 1984; Luz, Kolodny, and Kovach 1984). This can be broadly used to identify population mobility if this does not match the expected values for the environment in which they are buried (Pederzani and Britton 2019).
We measured the carbonate oxygen isotope ratios (δ18OCO3) of 33 second premolars from 17 individuals (1 Early Iron Age, 1 Middle Iron Age, 15 Late Iron Age/Early Roman) from Duxford and 15 individuals (all Mid-Late Roman) from Vicar’s Farm, Cambridge. We compared the results to published δ18OCO3 data from 33 individuals from Knobb’s Farm (1 Middle Iron Age, 32 Late Roman: (Wiseman et al. 2021). The datasets are broadly comparable, although due to the variety of teeth analysed, the data will not represent exactly the same period of life (Lightfoot in Wiseman et al. 2021, p. 160).
The δ18OCO3 values across the three sites are wide-ranging and overlapping (Figure 5) (Data S7). The distribution of the δ18O data slightly differs. However, the range of values seen at Duxford and Vicar’s Farm fit well within the distribution of the data from Knobb’s Farm (Figure 5).
|
|
|
Post by Admin on Dec 1, 2023 20:52:41 GMT
Mobility through isotopic analysis To further explore potential childhood origins and geographical mobility, we measured oxygen isotope ratios in the tooth enamel of Iron Age and Roman individuals excavated from the Cambridgeshire region. The oxygen isotope composition of local water sources is largely determined by the local climatic conditions (Dansgaard 1964; Pederzani and Britton 2019). The oxygen isotope ratios measured in archaeological human tooth enamel are a reflection of the water consumed during the formation of the enamel during childhood and can therefore provide information about the environment the individual grew up in (DeNiro and Epstein 1978; Longinelli 1984; Luz, Kolodny, and Horowitz 1984; Luz, Kolodny, and Kovach 1984). This can be broadly used to identify population mobility if this does not match the expected values for the environment in which they are buried (Pederzani and Britton 2019).
We measured the carbonate oxygen isotope ratios (δ18OCO3) of 33 second premolars from 17 individuals (1 Early Iron Age, 1 Middle Iron Age, 15 Late Iron Age/Early Roman) from Duxford and 15 individuals (all Mid-Late Roman) from Vicar’s Farm, Cambridge. We compared the results to published δ18OCO3 data from 33 individuals from Knobb’s Farm (1 Middle Iron Age, 32 Late Roman: (Wiseman et al. 2021). The datasets are broadly comparable, although due to the variety of teeth analysed, the data will not represent exactly the same period of life (Lightfoot in Wiseman et al. 2021, p. 160).
The δ18OCO3 values across the three sites are wide-ranging and overlapping (Figure 5) (Data S7). The distribution of the δ18O data slightly differs. However, the range of values seen at Duxford and Vicar’s Farm fit well within the distribution of the data from Knobb’s Farm (Figure 5).
|
|
|
Post by Admin on Dec 3, 2023 18:21:11 GMT
Figure 5: Raincloud plot of δ18OCO3 values from Duxford, Vicars Farm and Knobbs Farm, showing probability distribution, median, IQR, outliers and scatter of data, with individual skeleton numbers. Data for Knobb’s Farm sourced from Wiseman et al. (2021). Converting the δ18OCO3 values to phosphate oxygen isotope values (δ18OPO4) (Coplen 1988; Chenery et al. 2012) allows for broad comparisons with previously published data and expected ‘local’ environmental values (Figure 5). The mean expected ‘local’ range of δ18OPO4 values for the eastern low rainfall zone, in which the three sites are located, has been estimated at 17.2‰ ±1.3 (2SD) (Evans et al. 2012). The isotope values for Knobb’s Farm (mean: 17.2‰ ±2.2) fits well with this estimate, while the values for Duxford (mean: 16.5‰ ±1.6) and Vicars Farm (mean: 16.2‰ ±2.6) are slightly lower but are still within the general range. Only skeleton 2004 (δ18OPO4 = 14.8‰), a Mid-Late Roman male from Vicar’s Farm and skeletons 324 (δ18OPO4 = 14.8‰), a Late Roman male (genetics = XY, macroscopic sex=?female) and 1392 (δ18OPO4 = 19.1‰), another Late Roman male, from Knobb’s Farm have δ18OPO4 values which are on the edge or beyond the overall total range of ‘local’ values currently estimated for Britain (Evans et al. 2012, Lightfoot and O’Connell 2016). We investigated the presence of potential outliers further, following Lightfoot and O’Connell (2016) (Table 3)). The 1.5IQR method is considered most robust in this instance and identifies outliers only at Vicar’s Farm: skeletons 2028, 2034 and 2055. However, these are well within the overall range of values seen at Knobb’s Farm and may only appear as outliers due to small sample sizes. Interestingly, the individuals identified as outliers using 1.5IQR are not the same as those highlighted above as being outside the expected δ18OPO4 range, highlighting the difficulties of robustly identifying and interpreting potential childhood origins and geographical mobility using oxygen isotopes alone. Statistical comparisons of all sampled individuals from the three sites indicate that the samples were unlikely to be taken from populations with the same distributions (Kruskal-Wallis δ18OCO3: p=0.005, es=0.139), with the differences lying between Vicars Farm and Knobb’s Farm (Dunn’s post-hoc with Bonferroni adj: p-adj=0.007). For individuals that were confidently assigned a sex estimate of female or male, when both sex and site are considered, sex does not appear to correlate with δ18OCO3 values, but site does (Two-way ANOVA δ18OCO3 (site): p=0.033, es=0.119; δ18OCO3: p=0.803, es=0.001). There also appears to be no difference in the populations by time period when the individuals were assigned to broad date categories of Iron Age (incorporating those dated Early and Mid-Iron Age), Late Iron Age-Early Roman, and Roman (incorporating those dated to Mid-Late and Late Roman), (Kruskal-Wallis δ18OCO3: p=0.516, es=-0.010).
|
|
|
Post by Admin on Dec 6, 2023 20:31:02 GMT
Discussion The region of Cambridgeshire appears to have been composed mostly of homogenous, local populations when compared to the York samples of the same period where one out of seven randomly sampled individuals was a long-distance migrant. This evidence supports the hypothesis that the rural populations of Roman Britain were largely unaffected by migration. The differences between PCA mapping of Roman-period populations closer to Celtic populations and PiC score-based analyses show more similar regional affinities across different British geographic regions to Cambridgeshire Roman Period. This may potentially be explained by the relative homogeneity of the allele frequency pool before Anglo-Saxon migrations that makes the allele frequency approach more sensitive to detect the differences.
Regardless of relatively small sample sizes, we find close pairs of relatives in most of the Cambridgeshire sites. The exception being Knobb’s Farm, a cemetery associated with a settlement that was possibly engaged in the processing of agricultural products and in which there were a significant number of burials that were decapitations. Knobb’s Farm appears to have been more broadly networked than the other local farming communities sampled here, yet distinct from the cosmopolitan urban centre at York, which may explain the difference in population heterogeneity.
Generally, the oxygen isotope range of individuals from the three sites is wide, but falls within the range of previous archaeological data from the UK (Lightfoot and O’Connell 2016) and from Cambridgeshire specifically (Rose 2020). The δ18OPO4 values from all three sites are broadly within those expected for the British eastern bioclimatic zone (Evans et al. 2012), indicating that the majority of the individuals studied could have spent their childhoods in the local area, or at least, an area with similar climatic conditions to Cambridgeshire. Six individuals out of 33 can be considered different to the others. Three individuals are different to the rest in terms of the expected values for this region: skeleton 2004 from Vicar’s Farm and skeletons 324 and 1392 from Knobb’s Farm exhibit δ18OPO4 values which fall above or below expected ‘local’ values (Evans et al. 2012; Lightfoot and O’Connell 2016; Wiseman et al. 2021) which could indicate that they spent their childhoods outside Cambridgeshire (Table 3). These may be the most likely candidates for being ‘non-locals’ – skeletons 2004 and 324 may have spent their childhoods somewhere with a colder climate than Cambridgeshire and skeleton 1392 may have spent their childhood in a warmer, drier environment (Wiseman et al. 2021).
Three others, skeletons 2028, 2034 and 2055 from Vicar’s Farm, are statistical outliers for δ18OCO3 compared to the rest. This may indicate that these three individuals spent time in a geographically/climatically distinct area to the rest of the group when their teeth were developing. This could be particularly interesting as aDNA analysis identified skeletons 2028 and 2076 as likely to be brothers, however, their δ18OCO3 values are very different (skeleton 2028 = -3.20‰, skeleton 2076= -6.83‰), with skeleton 2028 having the highest δ18OCO3 value at Vicar’s Farm. This could indicate that the brothers were not raised in the same geographical location. However, the overall range of the Vicar’s Farm δ18OCO3 values is very similar to the other two sites and it is quite possible that the apparent bimodality is a by-product of the small sample size, and that if a larger number of samples had been analysed from the site, the distribution would be more normal and the difference between the brothers could be considered part of ‘normal variation’ at the site.
Overall, the results of the oxygen isotope analysis of the individuals from Duxford, Vicar’s Farm and Knobb’s Farm are fairly homogeneous and indicate that the majority of the population were likely to have spent their childhoods either in the local area, or in an area with similar climatic conditions to Cambridgeshire. There is no correlation between the oxygen isotope results and key demographic or burial information. There is no difference in oxygen isotope values between females and males at each site nor by time period, meaning there is no isotopic evidence for a large-scale change in population diversity (i.e. differences in the number of individuals who may have originated from elsewhere) between the Iron Age and Roman period sites. The data does appear to highlight several individuals that may have spent their childhoods elsewhere, however, the difficulties of interpreting δ18O data due to large ranges and broad, overlapping estimations of expected values means that these individuals can only be tentatively identified as potential migrants and further corroborating evidence, such as strontium isotope analysis would be required to make a more definitive interpretation.
|
|