French Neolithic discontinuities

Marie-France Deguilloux and colleagues Deguilloux:2010 present a short analysis of ancient mtDNA recovered from a Neolithic burial at Prissé-la-Charrière, between the Loire and Garonne valleys of western France.

The mtDNA sample in the end was only three individuals -- one haplogroup X2, one U5a and one N1a. Each is intriguing, as far as a single sequence can be, because all are rare or absent from France today. I think one shouldn't go far interpreting three samples, but they contribute to the view that Neolithic mitochondrial variation in Europe was very different from recent Europeans. The N1a and U5b sequences fit within the already-known Neolithic (and for U5a, Mesolithic) variation in central and northern Europe.

It is from the U5a that Deguilloux and colleagues make a point about possible Mesolithic population continuity.

Subhaplogroup U5b has also been encountered in German Neolithic remains from the Corded Ware Culture (Haak et al., 2008) and in the hunter-gatherers studied by Bramanti et al. (2009), although in both instances, the branches concerned were distinct from the U5b in the Priss sample. It is, however, worth noting that haplogroup U5 has been encountered in surprising frequency in the hunter-gatherers studied by Bramanti et al. (2009) and could correspond to a Mesolithic heritage.

The story of N1a is that it was very common in the central European Neolithic, even though it is very rare today. That was first noted by Wolfgang Haak and colleagues Haak:2005, and has in subsequent years been joined by the observation that the pre-Neolithic hunter-gatherers had yet other common haplogroups. The population history of Europe was a lot more interesting than we suspected 10 years ago.

Deguilloux and colleagues attempt a conservative explanation for the frequencies of N1a in Neolithic samples:

The widespread distribution of the N1a lineage in Early and Middle Neolithic northwestern Europe may indicate genetic continuity from Mesolithic populations. This scenario would support a Mesolithic contribution to the earliest Neolithic of Atlantic Europe. This would imply that the N1a lineage was already common in indigenous north European populations and that the spread of the Neolithic was principally the result of cultural diffusion. Although so far the N1a lineage has not been encountered among late European hunter-gatherers in central and north Europe (Bramanti et al., 2009; Malmstrm et al., 2009), it is worth noting that less than half of the hunter-gatherers' paleogenetic data come indeed from the pre-Neolithic period (predating LBK expansion). Finally, no paleogenetic data currently exist for the Mesolithic period in Western Europe. This prevents any conclusion being drawn about N1a occurrence during the Mesolithic period in those regions.

I will note this -- the more that N1a is replicated across the Neolithic of Europe, the less and less likely that its subsequent vast reduction in frequency could result from genetic drift. When there was only one or two samples from Central Europe with high N1a, it was at least possible that this was a local founder population that did not spread its mtDNA diversity very far. If it were localized, even in the central Danube (a fairly big region) it might be possible to maintain that the later decline of N1a to its present low frequency had been due to population replacement.

Now N1a seems like a real marker of the LBK, spread widely into Western Europe. It may be, as Deguilloux and colleagues suggest, that it will be found at substantial frequencies in earlier samples somewhere in Europe. We do want some explanation for how it got to be common in this culture area.

Dienekes has written about the study. His point is a good one: If N1a were present somewhere in pre-Neolithic Europe, it would require some kind of "partition" of the pre-Neolithic population, along with its propagation -- presumably southeastward -- into the LBK of central Europe. Seems doubtful.

The study includes an illuminating paragraph about the sources of contaminating sequence in these Neolithic extractions.

Strict precautions were followed during all procedures (including precautions during excavation) and proved to be effective, because all researchers who directly participated in this study (from people working in the field to those working in the laboratory) were genotyped and their sequences were never observed during analyses. However, European sequences were randomly found in clones (28% of the sequences obtained). These specific sequences are regularly observed in the laboratory, whatever the project tackled (including samples from Polynesia or South America), in clones from samples or negative controls. They are not reproducible for a specific sample and are different from researchers' sequences. These facts lead us to suspect the contamination of PCR reagents (Leonard et al., 2007). It was relatively easy, however, to discard those contaminating sequences from our analyses because they were largely in the minority when compared with endogenous sequences.

It would not be very difficult to compare the results from different labs and do a forensic-quality analysis of these reagent contamination events. Surely a good fraction of ancient DNA results prior to the last few years must represent such contamination. Nowadays people have the expectation that Neolithic-era remains may have rare or exotic haplogroups, but it hasn't been so long since people assumed that French equals French. I expressed some concern about this criterion before -- "strange" stands in for "non-contaminated" in too many studies.

It might be very helpful to have a paper outlining the actual contamination pathways that have been found to affect multiple labs. Then the results could be compared against reports that have come out over the years. If people are reluctant to cull doubtful ancient DNA results, at the very least they can target a set for replication studies.