Sunday, February 12, 2012

Is Neanderthal Your Distant Cousin or Your Ancestor?

The answer to the question is "mostly cousin" or "both," depending on whether you are African, Asian, or European. Lots of evidence suggests that people who migrated out of Africa in the past few hundred thousand years met and mated with indigenous populations of other humans who were already living in Asia and Europe.

The availability of many genome sequences coupled with our sophisticated understanding of population genetics allows workers to estimate how much Neanderthal DNA entered the modern European and Asian populations. It won't be long before similar studies are done with the Densovan genome.

There's lots of discussion about whether the strict "Out-of-Africa" scenario is still valid [Out of who knows where] [The scientists behind Mitochondrial Eve tell us about the "lucky mother" who changed human evolution forever]. There's little doubt that some version of the multiregional hypothesis is correct although it may not be as thorough as the original proponents argued.

John Hawks is an expert on this sort of thing and he's just posted some of his work on his blog [Which population in the 1000 Genomes Project samples has the most Neandertal similarity?]. He doesn't allow comments so you can ask questions here—he's been known to read Sandwalk when he gets bored doing real science.

I'll ask the first questions. John, what percentage of the genomes of Africans, Europeans, and Chinese are derived from the Neanderthal population? Your figure shows the amount of Neanderthal (Vi33.16) intrusion as something called "shared derived variants" but how much of the total genome does that represent? Is any of it in parts of the genome where there are genes?

[Image Credit: I don't know where this picture came from. I got it on Just Another Brooklyn Blog.]


  1. how much of the total genome does that represent?

    I'd like to know the answer to this both on an individual basis (my understanding is that each individual person is ~2.5-3% Neandertal-derived), and the related question of what fraction of the human genome has Neandertal alleles segregating?

    Is that 2.5-3% Neandertal ancestry largely the same between people, or do we all carry a different 2.5-3% of Neandertal within us? Given infinite breeding time, a carefully chosen starting population and no ethics whatsoever, could we in principle re-create a Neanderthal person from the buried fragments in modern human genomes, or are crucial parts missing?

  2. Thanks, Larry!

    The best point estimate for Neandertal ancestry for people outside Africa is 2.5 percent of the genome. To date, this has assumed that the value within Africa is zero. We can show now that some Africans have more than others, so the baseline value should be more than zero within Africa. That should elevate the estimate for people outside Africa but we do not yet know how much. Probably less than one percent, for a total less than 3.5 percent.

    The problem is analogous to determining how much of the genetic similarity between me and my grandmother is attributable to the fact she is my ancestor. We all know that I inherited roughly 1/4 of my genome from her (more or less). But the other 3/4 of my genome is very similar to hers also, because she was a distant cousin of my other grandparents and of every living person. If I didn't know that my grandmother accounts for 1/4 of my ancestry, I would compare how much more my genome shares with hers, compared to how much randomly chosen people share with hers. In this case, my answer depends to a considerable extent on what I assume about the genealogical patterns connecting my grandmother with those randomly chosen people. With my grandmother I'll probably get it close to right, but with a great-great-great-grandparent, my answer is quite sensitive to the assumptions.

    That's what we attempt to do with the Neandertal genome. Thus far, others have assumed a model in which Neandertals and Africans became isolated around 300,000 years ago, and have estimated the extent of Neandertal ancestry on that basis. That generates an estimate of 2.5 percent. I doubt the population model, and so I am more cautious about this. At present we are working to characterize exactly which gene regions come from Neandertals.

    Many of the Neandertal shared alleles are in coding parts of genes. We have been comparing the incidence of this sharing in coding and conserved intervals of the genome and I'll post on those soon.

  3. Thanks, for interesting posts!
    I would like to ask How we can know whether a variation is specific for Neadertals (appeared after specification) or whether it is common for both humans and Neandertal (inherited from common ancestor) if we don't know the full extent of variability of human (especially African) genomes?

    I am under the impression that people outside the Africa have very homogeneous genomes due to many genetic bottlenecks that has happen since their ancestors left from Africa. Their sequences provide very little information about human genetic variation.

    Much more human variation can found from African genomes. For example, Schuster et al. [1] has found that "In terms of nucleotide substitutions, the Bushmen seem to be, on average, more different from each other than, for example, a European and an Asian". As far as I know very few African genomes are available even from 1000 Genomes project.

    [1] Schuster et al. Complete Khoisan and Bantu genomes from southern Africa. Nature. 2010 Feb 18;463(7283):943-7.