Between the millennia and generations

When I first realized of the possible utility of genetics toward anthropology and history I came up against a major problem in addressing extremely fine-grained questions: the tools did not have the power to probe very small genetic distances, constrained in time and space. After reading Norman Davies massive The Isles I was to understand that the Anglo-Saxon transformation of Britain into England was a matter of elite cultural emulation. What did genetics have to say about this? The answer was “not much.” In an earlier era of a few dozen classical autosomal markers or just mtDNA and Y chromosomal fast evolving microsatellites genetics was not powerful enough to distinguish between extremely close European populations. In The History and Geography of Human Genes L. L. Cavalli-Sforza notes that Europe is the most homogeneous of his continental groups where he had many samples. If you remove outliers like the Finns and Sicilians, this is very true today when we have whole genomes. Though conventional whole-genome SNP panels can distinguish Northern European population clusters they are packed very tightly together.

Using the Human Origins Array data set the authors of Iron Age and Anglo-Saxon genomes from East England reveal British migration history projected the genotypes of ancient samples from the Iron Age and the Anglo-Saxon period upon modern groups. As you can see the Northwest European data sets are packed so close it’s hard to discern structure. In the era of pre-genomic phylogeography it was nearly impossible. To answer whether the Anglo-Saxon migration was a mass folk wandering or a tiny mercenary elite one would have to distinguish Germano-Scandinavian heritage from that of the native British, and the reality is that these are all very close groups. We now know that the ethnogenesis, at least in a genetic sense, of Northern Europeans as a whole seems to date to the Bronze Age, on the order of 4,000 years ago. Therefore, to pick up genetic structure you are by and large focusing on only two to three thousand years of drift between these groups! This explains why the older methods were so under-powered. There just wasn’t much raw material for them to work with; not enough time had passed for the populations to diverge.

A few months ago the PoBI project finally published, and revealed local structure across the British Isles. The authors concluded that in England proper a minor, but substantial, proportion of the ancestry derived from the Anglo-Saxons. That is, Germans. The authors used SNP data sets. That is, less than a million markers in the genome (and out of the twenty to thirty million SNPs in the human genome). But, they had a massive sample coverage from which to extract insights from.

This preprint takes a different tack. They take whole genome sequences from ancient subfossils. What they lack in sample size, they make up for in marker set (some of the coverage is high enough that they probably got decent calls), as well as the fact that they are sampling individuals who are from the source cultures in question. Celts, Anglo-Saxons during the invasion period, and also during a some what later epoch. Their results in nearly perfectly in line with those of PoBI. Using their samples as representative they estimate that ~30 percent of the eastern English ancestry is German. A lower proportion is found in Wales and Scotland.

The figure above illustrates this finding. A key point about this paper is to emphasize that since they are using whole genome sequences they can focus on rare alleles. Because the alleles are at low frequency they’re likely to be younger, and if they are younger they are also more likely to yield the power to discern differences between very genetically close groups, such as the Germans and the British. In the plot above you see that the Anglo-Saxon samples are shifted right, while the Iron Age samples are shifted to the left. What this shows is that the Iron Age samples tend to share more rare alleles with the Spanish IBS data set from the thousand genomes, while the Anglo-Saxons share more alleles with the Dutch.  Eastern England and Wales and Scotland occupy positions that you’d expect.

Furthermore, the authors utilized the program rarecoal to explicitly model the population history of Britain using their data. This will be a major feature of future work in this area, as researchers drill-down to a very fine grain and ask precise questions which good quality whole genomes and robust phylogenetic packages can actually tackle.

There is still much that needs to be worked out on this topic. There’s only so much you can say from these handful of individuals. But even with these finite samples much was extracted. The authors observe that one of their Anglo-Saxon era individuals, buried in an Anglo-Saxon fashion, clustered perfectly with the British Iron Age individuals. Additionally, this individual was outfitted in a manner which suggested they were very high status within Anglo-Saxon society. But the authors did not connect this with the fact that all their Anglo-Saxon individuals were female. Hypergamy is entirely typical in human societies, and it is plausible that large numbers of migrating German men arrived on British shores without a wife and family in tow. In the years after the Norman invasion it was not uncommon for noble Saxon houses to give their daughters to an invader. And so the Anglo-Norman aristocracy arose as a synthesis between distinct paternal and maternal lineages. A similar scenario likely played out during the invasions of the Dark Ages.