Models are great, because rejection is easy

By Razib Khan | April 23, 2013 1:20 am

There’s a new paper in PLoS ONE, Female and Male Perspectives on the Neolithic Transition in Europe: Clues from Ancient and Modern Genetic Data, which uses a combination of contemporary and ancient (that is, from subfossils) Y and mitochondrial DNA to understand the demographic past of Europe. Recall that the Y traces the direct male lineage, and the mtDNA the direct female lineage. Because they don’t recombine and generate clean converges back to a last common ancestor (there is no reticulation because there is no sex on these loci; they’re inherited from one of the two parents), they’re amenable to a lot of nifty demographic inference generation. In this paper they test specific models, and produce probability distributions of those models. Since it is open access I invite you to read the paper. The problem with these sorts of papers is I have a hard time trusting them until I replicate the results or have a sense of how cranky the software/code is!

But there’s a bigger problem. The authors find that ~80% of Spanish ancestry is “Paleolithic” and ~100% of that of Southeast Europeans is “Neolithic.” I don’t have a problem with these figures so much, except that autosomal DNA generally implies the major genetic cline in Europe is north-south, not east-west! By this, I mean that there is more variation between Poland and Greece, than Greece and Spain, and Spain and England, than Spain and Greece. At least on the whole genome scale. But this tends not to be the case when you look at Y chromosomes, for example. There’s an east-west split. Why? I have no idea, but the fact of this contradiction exists should make you cautious.

Finally, I have to throw in that it seems that this model is focused on one expansion of farmers into Europe. Though reading structure plots can be like reading tea-leaves, I think it is important to enter into the record that this may not be correct, and there may have been multiple agricultural expansions. If so, I wonder about the validity of the inferences if the assumptions are faulty. These are all things one has to consider when one tries to gauge the plausibility of abstruse statistical papers like this dependent on silico

Citation: Rasteiro R, Chikhi L (2013) Female and Male Perspectives on the Neolithic Transition in Europe: Clues from Ancient and Modern Genetic Data. PLoS ONE 8(4): e60944. doi:10.1371/journal.pone.0060944

MORE ABOUT: Human Genetics
  • GuestOfGuests

    “Though reading structure plots can be like reading tea-leaves”

    Very true.

  • Charles Nydorf

    Why base a study solely on uniparental markers now that autosomal data is available? This should only have been published if the model yielded plausible results and the model does not do this.

    • razibkhan

      1) lots of ancient DNA (especially mtDNA)

      2) lots of the explicit models are easier to test with uniparental (e.g., ABC)

  • ohwilleke

    I wonder if the W-E faultline in uniparental markers v. N-S faultline in autosomal genetics could be a product of relative admixture proportions (which are dominant in the uniparental markers that don’t really distinguish statistically based upon how deeply different the haplogroups are from each other) v. intensity of the difference which does show up in autosomal.

    Proto-Uralic populations should have some East Eurasian genes that are much more distantly separated from all other West Eurasians (at time depths approaching 70kya) than populations ancestral to W. Europeans are from populations ancestral to E. Europeans both of which are almost exclusively West Eurasian in ancestry (which are distinct at time depths of probably around 20kya or less). Considerably less Proto-Uralic admixture could produce comparable levels of variation in the sample.

    In other words, most European populations are mostly an admixture of populations ancestral to West Europeans and populations ancestral to East Europeans, but TMRCA of both W. Europeans and E. Europeans is fairly recent relative to the TMCRA of Proto-Uralic peoples and the common ancestor of W Europeans and E. Europeans. The intensity of the N-S genetic differences may amplify these differences despite making up a relatively modest share of admixture in current populations relative to W Europeans v. E. Europeans in a way that isn’t reflected in fairly basis comparisons of uniparental marker haplogroup variation. If the intensity of the N-S difference is sufficiently greater than its shortfall in admixture contribution, then N-S comes up as the first dimension of a PCA for autosomal genetics, while W-E comes up as the first dimension of PCA for uniparental genetics (since the order of dimensions in a PCA determined by the percentage of variation that they explain).


Discover's Newsletter

Sign up to get the latest science news delivered weekly right to your inbox!

Gene Expression

This blog is about evolution, genetics, genomics and their interstices. Please beware that comments are aggressively moderated. Uncivil or churlish comments will likely get you banned immediately, so make any contribution count!

About Razib Khan

I have degrees in biology and biochemistry, a passion for genetics, history, and philosophy, and shrimp is my favorite food. In relation to nationality I'm a American Northwesterner, in politics I'm a reactionary, and as for religion I have none (I'm an atheist). If you want to know more, see the links at


See More


RSS Razib’s Pinboard

Edifying books

Collapse bottom bar