Neolithic milk fog

Razib points today to an article in Der Spiegel about the revival of folk migration as an explanation for the Neolithic in Europe. His post (Vlkerwanderung back with a vengeance”) is worth reading. The general issues here are very interesting right now because the increase in data has made it possible to propose and test more and more complex scenarios. The simple scenario, gradual demic diffusion, appears wrong in many details. Archaeological cultures appeared and spread in spurts, which we now know were often composed of people genetically very different people.

The article in Der Speigel is titled, “How Middle Eastern Milk Farmers Conquered Europe”.

The main idea of the article is that our understanding of the spread of Neolithic cultures into Europe has been revolutionized by ancient DNA and more sophisticated chemical analysis of artifacts. That’s more or less correct. We really are thinking much more these days about folk migrations bringing new people into Europe. We know that lactase persistence was a recent evolutionary phenomenon in European groups, which was absent before the early Neolithic.

Problem is: from the standpoint of ancient DNA samples, the lactase persistence mutation was also absent within the early Neolithic! The article is full of details that are wrong or misleading. Most important, it links the appearance and proliferation of the lactase persistence trait with the LBK. This might appear to make sense. The chemical analyses have supported the importance of dairying and presumably milk consumption in the LBK. But the genes of the LBK skeletons don’t have the lactase persistence marker.

The absence of lactase persistence in these early Neolithic people is entirely to be expected. Such an allele couldn’t become common until the selection pressure was in place. People had to be drinking milk habitually at key times of vulnerability to establish this selection pressure. Even when the selection pressure is very strong, as it was for lactase persistence, the initial growth of a selected allele is very slow. It did not become common in Europe until thousands of years after it first appeared.

So lactase persistence did not distinguish early Neolithic people in Europe from agriculturalists in the Near East, because neither of those populations had it at any detectable frequency. All the stuff in the article about how lactase persistence originated in Central Europe? It’s irrelevant to whether these ancient populations were connected or not.

What does distinguish the early Neolithic in central Europe is the mitochondrial DNA. I’ve discussed this several times in the last few years (“Early European mtDNA: only mysterious if you want it to be”, and most recently “French Neolithic discontinuities”). The early Neolithic in Central Europe and France is characterized by several common haplogroups that are absent or rare in both earlier and later Europeans.

It remains to be seen whether we can document a clear analogue of this mtDNA observation with nuclear genetic data. We know a lot about the variation of present-day Europeans, but most attention to geographic relationships has been run through course filters – maps of the first two principal components are very striking in their correspondence to geography, but they really don’t address the timing of movements that may have contributed to the pattern.

The differences between early Neolithic and later Europeans suggests that post-Neolithic migrations – real Völkerwandurung – actually had a major impact on the European gene pool. What we see today is not a pattern established 6000 years ago, but a palimpsest richly painted with strokes from successive migrations.

One aspect of this scenario: There’s no reason to link the early Neolithic with Indo-European languages. There were many later widespread population movements that might have carried this language family, and we know that these later movements were genetically decisive – at least, as concerns the maternal genealogy. The relation of Y chromosome haplogroups with mtDNA haplogroups is a critical question, but even more necessary is the development of an effective means of testing these hypotheses with nuclear genotype data.