The Anglo-Saxon migration and the formation of the early English gene pool

Abstract

The history of the British Isles and Ireland is characterized by multiple periods of major cultural change, including the influential transformation after the end of Roman rule, which precipitated shifts in language, settlement patterns and material culture¹. The extent to which migration from continental Europe mediated these transitions is a matter of long-standing debate^2-4. Here we study genome-wide ancient DNA from 460 medieval northwestern Europeans-including 278 individuals from England-alongside archaeological data, to infer contemporary population dynamics. We identify a substantial increase of continental northern European ancestry in early medieval England, which is closely related to the early medieval and present-day inhabitants of Germany and Denmark, implying large-scale substantial migration across the North Sea into Britain during the Early Middle Ages. As a result, the individuals who we analysed from eastern England derived up to 76% of their ancestry from the continental North Sea zone, albeit with substantial regional variation and heterogeneity within sites. We show that women with immigrant ancestry were more often furnished with grave goods than women with local ancestry, whereas men with weapons were as likely not to be of immigrant ancestry. A comparison with present-day Britain indicates that subsequent demographic events reduced the fraction of continental northern European ancestry while introducing further ancestry components into the English gene pool, including substantial southwestern European ancestry most closely related to that seen in Iron Age France^5,6.

Fig. 1. Spatial and temporal origin of ancient individuals in this study.

a, Spatial distribution of sites analysed in this study. b, Temporal distribution of samples analysed in this study, with site occupancy ranges indicated by bars. Non-transparent symbols indicate radiocarbon-dated samples; transparent symbols are scattered uniformly along site occupancy ranges. DK, Denmark; ENG, England; GER, Germany, IE, Ireland; NED, Netherlands.

Fig. 2. PCA.

a, Present-day genomes from northwestern Europe. b, Published and novel ancient individuals in this study, projected onto a. Polygons indicate where two-thirds of the respective groups are located (England Bronze Age (BA) + Iron Age (IA) and North Sea IA + Early Middle Ages (EMA), respectively). The Scandinavian IA samples are connected with lines for clarity. For rough time boundaries of the samples used here, see Methods.

Fig. 3. Individual-based and site-based ancestry decomposition.

a, Individual supervised admixture results for Bronze Age (n = 140), Iron Age (n = 304) and Early Middle Age (n = 285) genomes from England. The symbols and colours of early medieval sites correspond to Fig. 1. b, Mean CNE and WBI ancestry estimates of British–Irish sites from the Bronze, Iron and Early Middle Ages.

Fig. 4. Identifying continental source regions for immigrant ancestry in early Medieval England.

Shown are (1) continental sites that are genetically indistinguishable from the more than 95% CNE EMA English (England EMA CNE) population using qpWave and provide fitting P values as source in a two-way qpAdm model of England EMA, as well as (2) the predicted locations for 72 England EMA CNE genomes using LOCATOR. The red dashed line marks where 95% of the qpAdm and qpWave data are located.

Fig. 5. Population structure of present-day Britain and Ireland.

a, Ternary plot of present-day British–Irish populations as a three-way admixture between late Iron Age and Roman England (England LIA Roman) (n = 32), France IA (n = 26) and England EMA CNE (n = 109). b, Boxplot comparison of France IA ancestry proportions in 23 English PoBI sampling regions using either England LIA Roman (n = 32) or Worth Matravers (n = 16) as source for local British ancestry in qpAdm. The P value obtained from a two-sided paired Student’s t-test is shown. The bounds of the box represent the 25th and 75th percentile, the centre represents the median, and the whiskers represent the minimum and maximum values in the data. Dashed lines connect points from the same region. c, Geographical distribution of the England EMA CNE, ancestries based on the interpolation of 31 present-day population estimates. The coordinates of the sample collection districts approximate the centroids of the averaged birthplaces of the grandparents. d, Same as c, but for France IA.

Extended Data Fig. 1. Genetic affinity statistics between ancient and present-day northwestern European populations.

a) F_ST between relevant present-day Europeans and England_EMA (n = 285). Populations are coloured to their respective language family affiliation. Belgium is classified as Germanic, since Belgian samples were recruited mainly amongst patients from the northern Flemish-speaking region of Belgium. Error bars represent ± 3 standard errors. Samples sizes for present-day European populations are indicated in Supplementary Table 3.1 b) F_ST between relevant present-day Europeans and England_IA (n = 290). Error bars represent ± 3 standard errors. c) F-statistics of the form F₄(YRI, TestA; Ireland, TestB). Negative values indicate that the test population is closer to Ireland than to TestB; positive values indicate that the test population is closer to TestB than to Ireland. d) Same for the F-statistics of the form F₄(YRI, TestA; Denmark, TestB).

Extended Data Fig. 2. Individual-based ancestry decomposition and population affinities through time.

a) f-statistics of the form F₄(YRI, Test; WBI, CNE) for 758 ancient English individuals. Data are presented as point estimates for the respective F₄-statistic ± 2 standard errors. Negative values indicate that the test individual is closer to WBI than to CNE; positive values indicate that the test population is closer to CNE than to WBI. b) Modelling ancient post-Neolithic individuals from England and Ireland as a mixture of CNE individuals (red) and the WBI individuals (blue). Data are presented as admixture proportions. Each bar represents genome-wide mixture proportions for one individual. Individuals are ordered chronologically.

Extended Data Fig. 3. Regional ancestry decomposition and population affinities through time.

a) for England: left) f-statistics of the form F₄(YRI, Test; WBI, CNE). Data are presented as exact F₄-values. Negative values indicate that the test population is closer to WBI than to CNE; positive values indicate that the test population is closer to CNE than to WBI. Error bars represent ± 2 standard errors. Samples sizes for Test populations are indicated in Supplementary Table 3.3 right) Mean CNE and WBI ancestry proportions per period as inferred using supervised ADMIXTURE. Error bars represent ± 1 standard error of the mean. b) same for Scotland. c) same for Wales. d) same for Ireland.

Extended Data Fig. 4. Family tree reconstruction featuring integration of local ancestry into an immigrant kin group.

a) The genetic pedigree of 13 related individuals at Dover Buckland. Indicated are the mtDNA haplogroups, Y-chromosome haplogroups, and associated grave goods of each individual. Males are depicted as squares, females as circles. b) Spatial distribution of the addressed burials across the site. Genetically related burials are connected with lines. c) Genetic distribution of the addressed individuals across a Principal Components Analysis of present-day genomes from northwestern Europe. Genetically related individuals are connected with lines.

Extended Data Fig. 5. Visualisations of genetic affinity between early medieval English individuals and contemporary continental populations.

a) Multidimensional Scaling Plot of pairwise F₃ distances of the form F₃(CHB, ancient population A, ancient population B). b) PCA of F₄ statistics of the form F₄(YRI, ancient population; TestA, TestB). TestA and TestB iterate through 15 present-day European populations (Methods). Additionally, a neighbour-joining tree of the same dataset was projected onto the two PCs. England_EMA_CNE are those early medieval English individuals who have exclusively CNE ancestry.

Extended Data Fig. 6. Measures of CNE ancestry in early medieval English sites.

a) F₄ statistics of the form F₄(YRI, Site; WBI, CNE). Data are presented as point estimates ± 2 standard errors. Samples sizes for early medieval English sites are indicated in Supplementary Table 5.3 b) Mean supervised ADMIXTURE proportions at K = 2. Error bars represent ± 1 standard error of the mean. c) qpAdm admixture proportions and p-values using a two-way admixture model of England_EMA with England_LIA_Roman (n = 32) and LowerSaxony_EMA (n = 39) as sources. Error bars represent ± 1 standard error.

Extended Data Fig. 7. Extended Principal Components Analysis.

a) Principal Components Analysis of present-day genomes from northwestern Europe. IE = Northern Ireland & Ireland, WA = Wales, SC = Scotland, ENG = England, NL = Netherlands, GER = Northern Germany, DK = Denmark, NO = Norway, SE = Sweden, BE = Belgium, FR = France. b) Genetic structure of published and novel ancient individuals in this study, projected onto a). Polygons indicate where 2/3 of the data is located (England BA+IA, North Sea IA+EMA, and France IA+WGermany EMA, respectively). The Scandinavian Iron Age samples are connected with lines for clarity. Western Germany comprises samples from Alt-Inden; Northern Germany comprises samples from Lower Saxony, Mecklenburg-Vorpommern, and Schleswig-Holstein; Denmark_MA comprises samples from Copenhagen. BA = Bronze Age (EBA, MBA, and LBA = early, middle, and late BA), IA = Iron Age, RA = Roman Age, MA = Middle Ages (EMA=early MA), For rough time boundaries of the samples used here, see Methods.

Extended Data Fig. 8. Correlations of early medieval ancestry with archaeological and linguistic evidence.

a) Distribution of early Anglo-Saxon cemeteries, Celtic river names, and Frankish objects. Ancestry proportions from supervised ADMIXTURE at K = 3 are shown for early Anglo-Saxon sites. CNE ancestry is shown in red, WBI in blue, and CWE in green. b) qpAdm ancestry proportions for 31 present-day populations from Britain and Ireland using LowerSaxony_EMA as a source for CNE ancestry. Number of individuals used for each source is indicated (England_LIA, LowerSaxony_EMA, and France_IA; n = 32, 39, and 26, respectively). Samples sizes for present-day PoBI sampling regions are indicated in Supplementary Table 5.16. Error bars represent ± 1 standard error.