Human evolution: a tangled bank

February 18, 2016 • 1:03 pm

by Matthew Cobb

Back in October, we looked at the discovery of anatomically modern human teeth in China, from 100,000 years ago. This was surprising because although archaeological evidence suggested that Homo sapiens first came out of Africa perhaps 125,000 years ago, it was thought that they hung around the Middle East, maybe venturing into Western Asia, but no further. Those teeth told us that we got further East than was thought.

This initial Out Of Africa event was generally thought to have ‘failed’, in that the genetic data from all modern non-African human populations suggest that our common ancestors walked out of Africa around 65,000 years ago. Everyone around the planet who is not from Africa is a descendant of that second wave of migration, which most certainly did succeed – and how.

There was no evidence that the first, ‘failed’ wave of migration left any genetic traces. Until now.

In a paper that has just appeared in Nature, researchers from around the world have studied the genome of a Neanderthal who lived in Denisova cave in the Altai mountains in Siberia over 50,000 years ago, along with genes from two Neanderthals from Spain and Croatia, and the genome of a member of the mysterious Denisovans, who lived in the same cave in Denisova (hence the name), although probably not at the same time as the Neanderthal. They also looked at modern African genomes, which should not have been affected by contact with Neanderthals or Denisovans, as their ancestors did not leave Africa, and at other modern genomes.

This study reveals quite how complex the interactions between these various forms of human were. Above all, they show that the Altai Neanderthal individual had inherited genes from a human who left Africa in that first wave of ‘failed’ migration, tens of thousands of years before the Altai Neanderthal was alive. Perhaps those first bold humans did not leave any genes in us directly, but they did leave genes in our close cousins, the Neanderthals of Siberia. Interestingly, there were no traces of such genetic mixing with the first wave of humans to be found either in the Spanish or Croatian Neanderthals, or in the Denisovan.

Finally, the Denisovan genome itself revealed that those people had not only swapped genes with the second wave of humans as they moved through Asia (this was already known), but also that, some time deep in the past, the Denisovans gained genes from an unknown source, mating with an individual or individuals who branched off from the rest of our lineage hundreds of thousands of years ago. (This final finding helps explain some mysterious results that suggested that some modern African DNA sequences seemed related to Denisovan sequences – in fact they shared DNA from one of our relatives in deep time.)

Here are two useful figures from the paper that help clarify this blizzard of astonishing facts. Firstly a figure showing the patterns of evolution and of swapping of genes between these groups. Time is from top to bottom, with the present at the bottom. Homo sapiens and their ancestors are in light blue, Neanderthals are in pink and the Denisovans are in red.

(c) Nature

You can see that the Neanderthal and Denisovan branches (in pink) fade away, showing that they disappeared (the red dots correspond to the samples the paper studied). The slanted lines at the very top are to show that that part of the figure is not to scale. The blue arrow going into the Denisovan is from the unknown forms that split off from the human lineage long ago, but after the chimpanzees. The red arrow going into the Altai Neanderthals shows the introgression from the first human migration, which on this figure stops at around 50,000 years ago, indicating that, for the moment, we do not think these people left any direct modern descendants. The exact relation of this group to modern Africans is not clear, and that’s why they’ve put the odd circle on the left-hand part of the figure.

Genetics is an amazing thing, and by comparing the frequencies of different forms of a given gene (these forms are called ‘alleles’), population geneticists can work out how many individuals must have been involved in producing that variability, in other words, how big the reproductive population was. This ‘effective population size’, as it is called, is much smaller than the demographic population – the actual numbers of individuals – but it gives you an idea of the general size. Here’s the same family tree, but this time with much a finer timeline, and, for each group, the effective population size as calculated using population genetics:

(c) Nature

This figure shows you that the effective population size of the Altai Neanderthals was really small – perhaps no more than 1000 individuals, compared to 27,000 for the Yoruba in Africa. The Altai mountains might have been a tough place to hang out, and the Altai Neanderthals, cohabiting in the region with the Denisovans, who appear to have had a larger effective population size, might have found life tough. The Altai individual’s parents were very closely related, perhaps half-siblings. The differences in the population sizes of modern humans and the Neanderthal/Denisovan group are striking. Maybe leaving Africa when they did was a bad idea – they do not seem to have flourished, compared to the humans who remained in Africa.

Here is a summary of our current knowledge in easy-to-digest bullet points. Cut them out and pin them to your bedside table. There will be a test in class next week.

  • Modern humans and Neanderthals diverged 550,000-765,000 years ago.
  • Neanderthals and Denisovans diverged 381,000-473,000 years ago.
  • Modern humans mated with Neanderthals in Asia around 100,000 years ago, during our first failed wave of migration, leaving our genes in their DNA.
  • Modern humans mated with Neanderthals in the middle East around 50,000 years ago, during the second successful wave of migration. We got some of their DNA, which has been linked with a number of characteristics.
  • Modern humans mated with Denisovans in Central and Eastern Asia, and got some cool genes from them, including one that enables Tibetans to live at high altitude.
  • Denisovans mated with an unknown form (NB we don’t know what the Denisovans looked like, beyond the fact that they had whopping teeth), perhaps a relict Homo erectus, the product of the very first wave of Homo migration from Africa

This is just the latest installment in what I think is the most astonishing part of modern biology. All of this would have been science fiction 20 years ago, simply because the data would have seemed impossible to access. Even once Svante Pääbo’s group began to sequence Neanderthal genes, no one could have imagined the discoveries that were to be made. Linking paleogenetics with archaeology and palentology is opening an amazing chapter in the history of our knowledge of ourselves.

Reference:  Kuhlwilm M, et al. (2016) Ancient gene flow from early modern humans into Eastern Neanderthals. Nature (in press)



63 thoughts on “Human evolution: a tangled bank

      1. Me, too. I’ve just now read it, a day later than it was posted. Boy, am I glad I went back and found it (I usually go back to read all those posts I’ve missed). Thanks, Matthew.

  1. With six introgressions (so far), three surviving to date, this is becoming interesting. Besides the intriguing possibility that the ghost lineage was H. erectus (with a reasonably old split with Denisovans at the time), my quick impressions were:

    – The initial Neanderthal/Denisovan (and possibly more) population could have been comparatively large to AMH. Even considering that the new data may have removed the 1-2 AMH bottlenecks of ~ 10 k and bumped it up to ~ 30 k.

    – The introgressions goes every which way. I used to think of the modern AMH as a “hardy hybrid” (due to acquiring immune system advantages from the Neanderthal introgressions), but if many or all archaics were too I guess I have to change my pity take. Modern humans as the “multitudinous mongrel” perhaps.

    In this context I just found a recent review of plausible (?) hominins. If I squint at their figure 2 of “more realistic estimate of the temporal ranges of the hominin species”, I can imagine that there were 7 contemporary human species and subspecies 100 kyrs ago. [“Hominin Taxic Diversity: Fact or Fantasy?”, Wood and Boyle, ; I haven’t had time to browse at that, just looking for that data.]

    1. I was looking for ideas that the mystery early crosses were H. erectus as well. I had thought that Denisovans might be H. erectus, but this chart suggests they are not old enough to support that thought.

  2. I assume the “French” and “Yoruba” of this diagram bear no special relation to the cheese-eating denizens of France or the Yoruba from Nigeria and Benin ?

  3. How is it that the data suggests that earlier H sapiens mated with Altai neanderthals? If those H. sapiens left no descendants how do we know of their DNA markers to look for them in the DNA of other extinct humans?

    1. Hmm, if the Denisovans have DNA related to some Africans, and this is used to suggest that Denisovans mated with a very early group (H. erectus?)..
      Then maybe the answer to my question is similar: that there is a smidgen of African DNA in this population of neanderthals, ergo an early wave of African H. sapiens migrated out, and left DNA in the Altai neanderthals.

  4. I write only to tell you how appreciative I am of this post.

    Some lines that leaped from the screen: “This study reveals how complex…” “Genetics is an amazing thing,” and “…figures that help clarify this blizzard of astonishing facts.”

    Thanks to researchers our long and complicated history is knowable (based on the best evidence available). Your clear explanation of findings like this increase my knowledge and add to my wonder about life but I ain’t taking your test!

  5. John Hawks has described hominid evolution
    with the analogy of a “braided stream” with
    many channels diverging and some then recombining through history.

  6. The complexity of our genetic history is astounding. Thanks for taking the time to distill these exciting discoveries in a way I could understand.

    1. quite fascinating. The scenario suggested here is that hominins (and probably hominids also?) regularly interbred with one another, even if at a low frequency. That is, Europeans are don’t have so much Neandertal DNA that one could imagine regular interbreeding which would have presumably produced a higher percentage of archaic DNA. I assume there is a way to estimate the rate of old time hanky panky.

      Does this analysis mean that there were no archaic humans in Africa? I would imagine that there was overlap of some kinds of hominins, even if other sub-species of homo sapiens were restricted to Europe or Asia.

      Also: did the common ancestor of Neandertals and modern humans most resemble modern humans? That’s my assumption, but I’d be interested in a comment by the intelligentsia.

  7. Great report Matthew.
    I was struck by the extremely low numbers of all of these populations. I can see there was quite a bit of luck in any group staying alive…well, until humans started breading like rabbits.

  8. Thanks for the easy-to-understand summary.

    We’re hearing about how Neanderthal sequences may be affecting us and it makes me wonder how our sequences affected them.

    If the Neanderthals (Ns) went extinct because they couldn’t compete with modern humans (MHs), it’s interesting that they weren’t rescued by interbreeding with MHs, that there aren’t now humans with mostly N sequences and others with mostly MH sequences. Does it suggest that MHs actually killed Ns?

    Also, it’s confusing to me that there aren’t a number of N sequences in MHs identified as being responsible for obvious adaptations to life in a northern climate – like light skin. One would think that such a trait, among many others, would have evolved in Ns and been passed on to northern MHs.

      1. Does there need to be a simple explanation?
        The simplest explanation amongst those equally supported by the evidence may be the Occam-ian choice. that doesn’t necessarily make it correct (though it does remain the way to bet).

  9. Excellent! I love the science posts, but not being a native english speaker I often have to read them more than once to fully understand them.

  10. I always have trouble with the certainty of some of the conclusions in studies of this sort. It seems to me that coalescent theory offers plausible alternatives to specific episodes of introgression and I have never understood how those are ruled out. A clear explanation from Matthew or a reader would be most welcome.

    For simplicity, lets imagine a population collectively carrying 1,000 copies of every autosomal locus. That establishes 1,000 gene lineages (at each locus) that will suffer the effects of drift (if nothing else) through the generations. At each generation, some lineages may be lost. Since loss is irreversible, at some point in the distant future, all but one of the original lineages will be gone (again, independently at each locus, and, of course, with much accumulation of younger sub-lineages). Looked at in reverse, this is the basis of the expectation that all modern day humans share a single common ancestor (at each locus) if we look far enough back in time. That is the coalescent. It is important to recognize that there is nothing whatsoever “special” about a last common ancestor thus defined. In particular, there is absolutely no reason to believe the coalescent corresponds in any way to the origin of modern humans.

    Now consider this process in a single widespread species. Different isolated populations will lose lineages independently, and this will take place randomly at various loci. Conversely, it is perfectly possible for two long-separated populations that differ at most loci to to have retained, purely by chance, the same lineages at certain sets of loci, even lineages that neither shares with other populations. So my basic question boils down to this: How would the “signature” of this sort of event differ from introgression between long-separated populations?

    1. If I understand correctly what you’re getting at, then it seems to me that the answer is that you’re ascribing far too much to chance. Functional proteins that do important work don’t just randomly disappear in a lineage. Even when they stop working due to some other change that makes then redundant, their genetic skeleton remains.

      In other words, culling of genes is not random. And some genes are too important to be culled at all. At most, they’ll be modified in either unimportant ways (DNA sequence substitution with no change in the protein’s amino acid sequence, or an amino acid sequence change that doesn’t alter the shape of the resulting protein enough to make any difference), or be changed by selection into something else – but still be recognizable as based on the same aboriginal protein.

      None of these ideas are foreign to geneticists. They choose, for comparison, genes which can be reasonably expected to yield useful information about relationships.

      The fact that relationships based on genetic analysis so often agrees with relationships based on morphological comparison is no accident.

      1. It’s not genes that are lost, it’s functionally equivalent but slightly different versions (alleles) at each locus.

  11. Thanks Matthew!
    I attended a seminar Wednesday by Christopher Walker, a paleoanthropologist who’s part of the Rising Star team working at the Dinaledi in South Africa.
    I suppose Homo naledi is older then any of the homonins named in this paper (no dates from Denaledi yet) but I wonder if they’ve gotten any DNA there. It’ll be interesting to see where they fit on that tree.

    Christopher works on homonin locomotion and had some very interesting thins to say about how H. naledi had legs very much like modern humans.

    1. I suppose Homo naledi is older then any of the homonins named in this paper (no dates from Denaledi yet)

      ISTR that Berger et al had the stratigraphy of the ossiferous deposit fairly clearly pinned by an over-laying “stal” (“flowstone”) layer to over a mega-year ago. I recall that they were looking for soot in roof-deposits of flowstone to try to anchor their stratigraphy better (much of the access route being infeasible for a single traverser carrying their own (fat+wick) lamp.

  12. Thanks for this interesting news, Matthew! It’s so great to have scientific discoveries explained with relative immediacy (compared to reading it in popular magazines in he old days) and articulated so well. Kudos!

    1. I agree completely. Just to note, how often do popular media even get the point, much less the details, right on scientific papers and especially those on evolution and human ancestry? Rarely! In mainstream media world, scientists are still looking for the “missing link” – though frequently the media gush that ZOMG scientists may have found it!

      Once I was exposed to public radio in college, I couldn’t stand to listen to or read mainstream (corporate) media any more – especially politics and economics, which is just horrible on those topics. Then I got into reading biology and WEIT, and that made even public media unbearable, except NOVA or course and one local (SF Bay) radio program hosted by a secular JewBu.

      There is nothing like the straight shot you get from PCC, Matthew et al.

  13. Homo sapiens first came out of Africa perhaps 125,000 years ago, it was thought that they hung around the Middle East, maybe venturing into Western Asia, but no further. Those teeth told us that we got further East than was thought.

    The presence of the Dimanisi (Georgia, Caucasus) Homoe erectus-ish have long reminded people that our knowledge of the dispersal of hominins into the non-African Old World is, at best, extremely fragmentary. There’s no good reason to think that our knowledge of the dispersal of AMHs and Neanderthalers is much better. Holes everywhere, and barely a bone to fill them!

  14. One of these days I am going to get my DNA ancestry done by one or two of the providers (among the hold-ups for me is making time to research which companies and which tests are worth doing). I wonder how long it takes for them to integrate findings like the one described here – assuming the data informs the ancestry DNA models; I don’t know enough to tell either way. I assume at least some of them will update subscribers – “Hey, we’ve narrowed your 5% Neanderthal heritage to within N timeframe and X,Y location!”

    An acquaintance did his Ph.D. in bioinformatics (or some such) 20-odd Yeats ago, and went on to help develop the chips that sped up the original Human Genome project: I am astonished at how rapidly that initial work was completed, and how quickly – and increasingly quickly – knowledge is expanding in the field.
    Perhaps that growth and its acceleration were predictable given Moore’s Law and the dynamics of data sharing and whatnot; predictable and still astounding. Thanks so much, Matthew!

  15. Thank you Mathew for this post. Several weeks back I saw a program on PBS (wish I could remember the title) on the topic of the role of interbreeding in evolution. We think of evolution leading to new species due to separation (often geographical) of populations. But we may ignore or downplay the role of interbreeding among populations thought to be separated. The program covered much of the same territory in Mathew’s post but concentrating on traits we may have evolved due to this interbreeding of modern humans with Neanderthals, Denisovans, and possibly Homo Erectus. This is why I like PBS, the fascinating science programs, many on Nova, on this network.

  16. [Note- I’ve had much trouble posting this – I hope it doesn’t end up in duplicate]

    Thanks for this splendid post, Matthew, one of the best and clearest I’ve read.

    I had coincidentally just read Svante Paabo’s fascinating memoir about these astounding discoveries. The thrill of the discoveries and the amazement he and his colleagues felt is excitingly recounted.

    This is accompanied by details of his career and private life, the struggles to obtain suitable samples from sometimes reluctant museums, the huge difficulties in avoiding contamination of samples, and the mixture of cooperation and sometimes rivalry with other scientists and laboratories.

    The book’s title “Neanderthal Man” actually doesn’t fully reflect his achievements. The emergence of the Denisovans as a separate group was a total shock. I can’t help wondering where his Nobel Prize is.

    Your post sets it all out in wonderful clarity.

    As always, more questions arise about the hominid family tree – eg, when if ever will we have DNA from Homo Floresiensis (difficult, given the tropical location); and of course from the Naledi discovery.

  17. So, looking at the timescale, is it the best guess that Neanderthals and Denisovans are the results of an earlier exodus (say 600ky), rather than the descendants of the even earlier Homo erectus exodus?

  18. Excellent article with an excellent title.

    As someone with only a very limited knowledge of genetics, I was easily able to understand the break-down of the paper. Also, as an interested non-biologist, I know Darwin used the term “tangled bank,” but it didn’t really make sense to me until now (it was still just a tree to me).

    And now I would like to make a request of knowledgeable commentors: Can anyone give me an easy to understand explanation of “effective population?” I wasn’t lazy, I looked it up on Wikipedia. The math made sense, but the article seemed to bounce around on the definitions of the terms.

    Thanks in advance

    1. I’ll give the definition from chapter 10 of Douglas J Futuyma, Evolution:
      The Effective population size (denoted Ne) of an actual population is the number of individuals in an ideal population (in which every adult reproduces) in which the rate of genetic drift (measured by the rate of decline in heterozygosity) would be the same as it is in the actual population. For instance, if we count 10,000 adults in the population, but only 1,000 of them successfully breed, genetic drift proceeds at the same rate as if the population size were 1,000,and that is the effective population size, Ne

      This makes sense, as clearly, for multicellular organisms, only those organisms that reproduce will get their genes into the next generation. And this is pretty much the Sewall Wright definition discussed above the contents box in the wikipedia article.

      The cases discussed below the contents box in the wikipedia article adjust for factors leading to such idealised populations being smaller than the census size, or factors that impact the ideal population size.

      Hope this helps

      1. Great, WordPress stripped my subscripts.

        I also forgot my quotation marks. Everything after the second Ne is my fault. Douglas Futuyma is innocent.:)

  19. “The royal bloodline isn’t what it used to be. Too much intermarrying I suppose. I always say: ‘When you reduce a family tree to a family bush, you just can’t hide as much beneath it.'”

    Neat stuff. It is all so *complicated*, but what else is new? 🙂

  20. Thank yo again, Mathew. Well done.
    That is a definite help in digesting this complexity. I’ve been trying to take in details, sorting them in my mind, and mostly being unclear about most. You have nicely deconfusionized this fellow curious explorer – a little bit. I understand better now than before reading your article, but, of course, now there are more questions to explore. The questions, at least, are becoming clearer.

    One thing that really teases me on a curiosity level, it that of what some of these different linesges looked like. We would all, I’m sure, like to have better visualization of who was who in the tangled migrations and genealogies. Can genetic search tools help us to better understand the developments of brains and cognitive functions in some of these ancient contemporaries? Is there any technological hope that we will some day be able to read more or reconstruct more from the information contained in genetic samples? Short of finding appropriate bone fossils or casts or artifacts, maybe I’m hoping for too much from our genetic sleuths.

    LOL… My curiosities relating to this started with How and Why Wonder Books when I was a kid. Now, as I find the time left to me running out, I am wanting to find ‘the answers’, and of course, I am still finding instead, more questions. That’s not so bad. We wouldn’t really ever want to run out of questions.
    “Confusion and curiosity rule together.” – me

  21. Lots of awesome info to digest. This is the kind of stuff on here I love to see…more, as available, please. Complicated, but not so much that a curious layperson can’t digest the best bits.

Leave a Reply