Denisovans are us

September 1, 2012 • 11:09 am

by Greg Mayer

One of the most exciting recent developments in paleoanthropology has been the discovery, based on just a few bones, of a previously unsuspected type of human called Denisovans (named for the cave where they were found in Siberia). Because of the paucity of skeletal evidence, our knowledge of them is based almost entirely on DNA extracted from the bones. Both Jerry and I have been following developments in these genetic studies here at WEIT (here, here, and here), and so the release of a new paper in Science by Matthias Meyer and a cast of thousands (well, 33, actually) (abstract only) is of great interest, presenting a high quality full genomic sequence of Denisova Woman.

Tree of 11 modern plus the Denisovan genomes, showing interbreeding event(s).

The paper’s results are well summarized in the above figure: in comparison with 11 geographically diverse modern Homo sapiens genomes, the Denisovan genome is distinctive, but shows evidence of interbreeding with Papuans, such that 6% of the Papuan genome derives from the Denisovans. Based on genetic evidence, they date the modern-Denisovan split at about 800,000 years ago (this would also be the date of the modern-Neanderthal split, since Neanderthals and Denisovans split from each other after the moderns split off), and date the Denisovans themselves to about 80,000 years ago.

This essentially confirms what had been found by Reich et al. (2010) in an earlier study (of which Meyer was a coauthor) of the same material, which had not achieved as high quality a determination of the Denisovan genome. The new study is also noteworthy for its technical advances in the genome sequencing of ancient DNA. Interestingly, modern humans from Asia (through which the ancestors of the Papuans had to pass on their way to Papua), do not show a Denisovan genetic influence, so the history of migration through Asia must involve at least two migration events (one in which the Papuans’ ancestors encountered and interbred with Denisovans, and another in which the ancestors of modern Asians did not).

Matthias Meyer at work in the clean lab at the Max Planck Institute. Credit: MPI for Evolutionary Anthropology.

The current work tends to confirm the conclusion that archaic humans (Neanderthals and Denisovans) were part of a group of interbreeding populations in nature that included the immediate ancestors of modern humans, and thus were members of the species Homo sapiens. This was a conclusion that Jerry, I, and, independently, John Hawks, the University of Wisconsin paleoanthroplogist, had reached. Hawks has a very interesting and detailed take on the new paper which is well worth reading.

In another recent paper (may be abstract only), Anders Eriksson and Andrea Manica argue that the shared components of the modern and archaic genomes may result from the retention of variation from an ancestral African population, and not from interbreeding outside of Africa; John Hawks finds their conclusion unlikely.


Eriksson, A. and  A. Manica. 2012. Effect of ancient population structure on the degree of polymorphism shared between modern human populations and ancient hominins. Proceedings of the National Academy of Sciences (USA) 109:13956-13960.

Meyer, M., et al. 2012. A high-coverage genome sequence from an archaic Denisovan individual. Science in press. [This is presumably a MS in press, but Science in the past has released preprints whose subsequent publication history has been murky- think arsenic based life.]

Reich, D., et al. 2010. Genetic history of an archaic hominin group from Denisova Cave in Siberia. Nature 468:1053-1060.

41 thoughts on “Denisovans are us

    1. The tree was constructed by an as yet unpublished method implemented in a program called TreeMix. The method is described in this preprint. Roughly, it models allele frequency change under drift with the possibility of migration (gene flow). The horizontal axis represents the amount of genetic drift that has occurred (thus, it’s 0 at the root, before they diverged). I’m not sure how we should think of this parameter in terms of units.


      1. You can think of those units as roughly being time in generations divided by [effective] population size. This is the “unit” of genetic drift. However, the branch lengths are only approximate as the Treemix authors make a number of approximations, see below eqn. 2 in the preprint. A newer version of the preprint is available on the arXiv. See also here for some discussion, comments on the TreeMix paper are welcome here

      2. I think the part I’m having the most trouble wrapping my brain around is the yellow line running from “way back” Denisovan to “not-so-way-back” Papuan. My error, I’m sure, is wanting to read x axis as a time measure, which as y’all point out, it isn’t exactly. Maybe it will sink in as I follow the comments.

        1. The yellow line indicates that the Papuans are modeled as admixed, with 6% of their ancestry more closely related to the Denisova, and 94% of their ancestry more closely related to modern humans. The x-axis approximately a measure of both time and population size, so longer branches mean more time an/or smaller populations. The Denisova had a small effective population size, and thus has a long branch.

      3. As Graham points out, the units are roughly t/2N, where t is generations and N is effective population size (if the population size changes over time, the interpretation is that it’s the integral over time of 1/2N dt). These sorts of units pop up all the time, for example, this is the diffusion approximation drift time as well.

        To get this to work in any reasonable time, we had to make some approximations; see the preprint for details.

  1. a new paper in Science by Matthias Meyer and a cast of thousands (well, 33, actually)

    The paper was coauthored, so it should read as Matthias Meyer and Martin Kircher.

    Now why do you suppose the author’s list has Jay Shendure in front of Jacob Kitzman, his grad student?

      1. Point being, you were writing facetiously when you said “and casts of thousands.” If I was the second co-author, who contributed equally as stated, I’d be annoyed at being the collateral of your generalization.

          1. & how do you tell which author contributed how much? What about the third named? The fourth?

            Citing only the first-named author before et al. was common practice in my experience. And we listed authors alphabetically, regardless of how much each had contributed (eg I was the “third” author of this).


          2. It’s a well-known convention in the biological sciences that the paper is the brainchild of the first author, or co-first authors. If they contributed equally, it may be alphabetical, but many people still consider the second co-author less of a contributor. That’s why they’re called lead authors. The last author is typically the PI.

            So on physics papers, first author is meaningless?

  2. So Adam and Eve mated and produced offspring, and some of those offspring mated with other similar-but-different hominids god created, and then the Flood was brought about to destroy all human life except for Noah and his immediate family, and all Denison and Neandertal (and any other non-human hominids) were put to death in the Flood. But.

    Noah or his wife (or both) was a descendent of a child produced by a mating of Denison and Homo Sapiens (with some Neandertal mixed in there too), so when the waters receded, and the Noah family inbred the beginnings of the present human race, DNA of these otherwise deceased creatures continued to be passed along.

    I suppose Creationists already have a story that covers Neandertals that I am unaware of. Perhaps I will become acquainted with it when the group deals gets around to rationalizing Denison and any other looming discoveries of hominids.

    Maybe I’ll hear their story about Pygmies then, too. I’ve managed to miss that one so far.

  3. Just for fun, I ordered the new Geno 2.0 kit from the Genographic Project. It tests the mitochondria or Y-DNA, as well as autosomal DNA, and as a fun bonus it tests for any Neanderthal or Denisovan DNA. I haven’t received the kit yet (I think it will be sent to me in October). I can’t wait to see if I have any Neanderthal or Denisovan DNA! That would be very cool!

    1. 23andMe has a similar kit. I’m proud to be a 3 percenter, a full two tenths of a percent more Neanderthal than my mom. I’m in the 95th percentile on this so my late dad must have been even more Neanderthal than I am.

      And I never noticed his brow ridges. 😉

  4. As an update to Jerry’s article on Pääbo’s views: “So I don’t agree with Svante Pääbo, who is quoted by Gibbons as saying “I think discussion of what is a species and what is a subspecies is a sterile academic endeavor””, he is still supporting population history as such but comes down the side of same species:

    “So does that mean that Denisovans were a separate species?

    “I really stay away from species designation,” Pääbo said. “I would not call it a different species, but clearly different groups with a different history. I would not call the Neanderthals a different species from humans either, actually.””

    Speaking of population history, don’t forget the unpublished but still arguable prerelease from Hawks on Ötzi as having twice as many neanderthal shared derived variants, and clearly outside today’s distribution. I’m a layman here, but wouldn’t that indicate some differentiation among subpopulations, especially if now the Denisovan results resolves asians as having more neanderthal shared variants?

    1. It will depend on the pattern of population contraction/expansion, local extinction and migration. Modern humans arrive in a first wave, mixing with the archaic population. A new migrant wave arrives replacing/mixing with the descendants of the previous modern human wave, diluting the archaic human contribution, etc.

      This could leave a somewhat patchy distribution of populations, with widely different levels of admixture. Hunter-gatherer population densities are very low, and neandertal and denisovan numbers were probably lower than what one would expect from more modern hunter-gatherers.

      I’m more interested in the fact that the denisovan reference sample is geographically very far from the routes taken by the modern populations showing the admixture. We need a sample from «local denisovans,» but then we were lucky in getting that finger.

      1. Thank you, a very clear picture of pertinent patterns.

        Maybe you saw this, but Hawks’ remark on the demography was:

        “That is a mystery to be explained. How did Asians end up lacking any evidence of Denisovan ancestry, when the peoples of Sahul (Australia and New Guinea) have six percent? It’s nutty!”

        And he goes on to later discuss possibilities as well as the immediate scope of better observations.

  5. I’m new to biology, so bear with me:

    Does this article say that modern humans, Neanderthals and Denisovans are all subspecies of Homo sapiens?

    1. It means that they are all variations within Homo sapiens. The concept of ‘subspecies’ refers to a particular geographic pattern of within species variation, namely that the variation is such that if you tell me where an individual is from I can tell you what it looks like, or, conversely, if you tell me what it looks like I can tell you where it’s from. The concept has sometimes been extended to variation in time as well (‘chronosubspecies’). Subspecies are also an “optional” taxonomic designation– you don’t have to name them. Neanderthal man and modern man have been given taxonomic names which would allow subspecies (or chronosubspecies) recognition, but there is a complex and imperfectly understood geographic and temporal sequence which would need to be sorted through before a comprehensive taxonomy (including a name for the Denisovans) would be advisable, in my opinion. You could in the meantime refer to Neanderthals as Homo sapiens neanderthalensis and moderns as Homo sapiens sapiens if you really wanted to use subspecific names.


  6. I don’t get it. Why should we think Denisovans and Neanderthals as members of H. sapiens? Interbreeding was rare and did not happen during most of those millenia that the populations coexisted.

    1. As I understand it, the definition of a species is a group of individuals who can produce viable offspring amongst themselves but not amongst others. By this definition they still constitute a species, even if the level of inbreeding was actually small.

      1. Dogs, wolves, and coyotes can all interbreed and produce viable young and their young can produce viable young. However, they are considered separate species. I think with genomics, the concept of species and subspecies is becoming a bit arbitrary and will eventually be defined by some kind of degrees of difference between populations. Perhaps?

        1. It’s a very good point, and yes there is a lot of grey area in it. I’ve also heard it pointed out that many dog breeds are of such different sizes that that they can’t physically mate. (How could a Great Dane mate with a chihuahua for example?) And therefore they could perhaps be considered different species.

    2. The full extent of interbreeding is not yet known. But if we reflect on recorded history, there are many instances of invading populations replacing the native population with little genetic exchange between populations. In some cases the gene flow is asymmetric (i.e. more genetic mixing into one population than the other). But there’s no question of interbreeding being fully possible, or that current modern humans constitute more than one species. As we find (I hope) more Denisovans, and more archaic genomes are sequenced, we may be able to answer some questions more fully, but the observed extent of genetic mixing is compatible (indeed, supportive) of conspecificity.


  7. Is it a common pattern in speciation that only one out of many similiar species (like H. Sapiens, Neandethal, Denisovan et al) persists in long time?

    Chimpanzees, Bonobos, Gorillas all seem to have found niches but other early humans seem to have been driven to extinction either passively or agressively leaving the world with a single species (plus a few fragments of DNA hidden therein.)

    1. I’m confused. The article argues that these are all one species, and that you have to use either chimps and bonobos as possible counterexamples (but you claim niche separation instead of geographical separation which is the usual claim), or go back in deep time to H. floresiensis or H. erectus vs H. habilis.

      Don’t you have many extant counter examples outside of hominins? Cichlids, squirrels, mice, rabbits et cetera.

  8. Can anyone speak to the issue of erecting a new taxon, H. denisovensis (?), on the basis of such scanty physical evidence? Is it possible that the remains found are simply those of a highly aberrant individual?

    1. Species have been named on the basis of DNA evidence more or less alone, but the only examples I can think of are ones where it was legally important to have a species name, i.e. there was a presumed endangered species for which the describers declined to offer more material in their descripiton because they thought collecting it would further endanger the species, but having a name enhances the prospect for legal protection. In the Denisova case, I think it would be wise to hold back on naming it till more material is found. There is a risk that some unscrupulous person (i.e not the people studying the finds) will publish a description (it’s happened before), but while this would be unfortunate, it would not truly impede scientific understanding.


      1. Good point, sometimes it is a matter of politics. There’s a north american wolf researcher in my lab so I’ll have to ask, but I remember the Eastern Timber/Canadian/Red Wolf being a good example of a controversial species, susb-species or wolf/coyote hybrid, depending on your political leaning.

  9. This was interesting –
    “We estimate that the proportion of Neandertal ancestry in Europe is 24% lower than in eastern Asia and South America (95% C.I. 12–36%). One possible explanation is that there were at least two independent Neandertal gene flow events into modern humans (18). An alternative explanation is a single Neandertal gene flow event fol-lowed by dilution of the Neandertal proportion in the ancestors of Euro-peans due to later migration out of Africa. However, this would require about 24% of the present-day European gene pool to be derived from African migrations subsequent to the Neandertal admixture.”

Leave a Reply to Keith Cancel reply

Your email address will not be published. Required fields are marked *