A recent study in Nature maps very fine regional variations in genetic make-up in the UK by applying more powerful statistical methods to whole-genome data originally acquired as part of a large-scale medical study (see bit.ly/1ETl1QC for the full article).
The results are very interesting indeed. Seventeen distinct groups emerged, each with an unambiguous regional connection. There is no single “Celtic” group. Rather, distinct clusters of common descent are clear across Cornwall and Devon, Western Scotland and (Northern) Ireland with, intriguingly, two very different groups appearing to populate North and South Wales. It’s even possible to make out a genetic reflection of the traditional link between the Fermanagh/ Tyrone area and the borders of England and Scotland.
The statistical methods used are highly sophisticated and persuasive. Most persuasive of all is the fact that these geographical groupings emerged from a blind analysis, with no pre-sorting into areas of where the samples were taken. There is no doubt that studies like this will be highly important evidence for pre-historic migrations, though the caveat remains that any DNA study is a snapshot of the present. Inferences about past migrations can only ever be extrapolation, and the present is rarely an accurate guide to the past.
Another reason for caution in this particular case is the size of the data. These consist of samples taken from just 2,039 individuals in the UK, statistically modelled against 6,209 individuals from across mainland Europe. However sophisticated the analysis, this sample population is so small that even a handful of genealogical outliers could well be skewing the results.
And some place called “Eire” is excluded from the study, on the basis that that initial analysis provided too much evidence of shared Irish ancestry with the UK, because “Eire acts as a source and a sink for ancestry from the UK”, perhaps reflecting millennia of two-way migration. Or shared ancestral populations. Or maybe both: “Which severely complicates interpretation of estimated ancestry profiles.”
Well, yes. Welcome to the severely complicated real world.