TED日本語 - スバンテ・ペーボ: DNAがつなぐ私たちの内なるネアンデルタール人

What I want to talk to you about is what we can learn from studying the genomes of living people and extinct humans. But before doing that, I just briefly want to remind you about what you already know: that our genomes, our genetic material, are stored in almost all cells in our bodies in chromosomes in the form of DNA, which is this famous double-helical molecule. And the genetic information is contained in the form of a sequence of four bases abbreviated with the letters A, T, C and G. And the information is there twice -- one on each strand -- which is important, because when new cells are formed, these strands come apart, new strands are synthesized with the old ones as templates in an almost perfect process.

But nothing, of course, in nature is totally perfect, so sometimes an error is made and a wrong letter is built in. And we can then see the result of such mutations when we compare DNA sequences among us here in the room, for example. If we compare my genome to the genome of you, approximately every 1,200,1,300 letters will differ between us. And these mutations accumulate approximately as a function of time. So if we add in a chimpanzee here, we will see more differences. Approximately one letter in a hundred will differ from a chimpanzee.

And if you're then interested in the history of a piece of DNA, or the whole genome, you can reconstruct the history of the DNA with those differences you observe. And generally we depict our ideas about this history in the form of trees like this. In this case, it's very simple. The two human DNA sequences go back to a common ancestor quite recently. Farther back is there one shared with chimpanzees. And because these mutations happen approximately as a function of time, you can transform these differences to estimates of time, where the two humans, typically, will share a common ancestor about half a million years ago, and with the chimpanzees, it will be in the order of five million years ago.

So what has now happened in the last few years is that there are account technologies around that allow you to see many, many pieces of DNA very quickly. So we can now, in a matter of hours, determine a whole human genome. Each of us, of course, contains two human genomes -- one from our mothers and one from our fathers. And they are around three billion such letters long. And we will find that the two genomes in me, or one genome of mine we want to use, will have about three million differences in the order of that. And what you can then also begin to do is to say, "How are these genetic differences distributed across the world?" And if you do that, you find a certain amount of genetic variation in Africa. And if you look outside Africa, you actually find less genetic variation. This is surprising, of course, because in the order of six to eight times fewer people live in Africa than outside Africa. Yet the people inside Africa have more genetic variation.

Moreover, almost all these genetic variants we see outside Africa have closely related DNA sequences that you find inside Africa. But if you look in Africa, there is a component of the genetic variation that has no close relatives outside. So a model to explain this is that a part of the African variation, but not all of it, [ has ] gone out and colonized the rest of the world. And together with the methods to date these genetic differences, this has led to the insight that modern humans -- humans that are essentially indistinguishable from you and me -- evolved in Africa, quite recently, between 100 and 200,000 years ago. And later, between 100 and 50,000 years ago or so, went out of Africa to colonize the rest of the world.

So what I often like to say is that, from a genomic perspective, we are all Africans. We either live inside Africa today, or in quite recent exile. Another consequence of this recent origin of modern humans is that genetic variants are generally distributed widely in the world, in many places, and they tend to vary as gradients, from a bird's-eye perspective at least. And since there are many genetic variants, and they have different such gradients, this means that if we determine a DNA sequence -- a genome from one individual -- we can quite accurately estimate where that person comes from, provided that its parents or grandparents haven't moved around too much.

But does this then mean, as many people tend to think, that there are huge genetic differences between groups of people -- on different continents, for example? Well we can begin to ask those questions also. There is, for example, a project that's underway to sequence a thousand individuals -- their genomes -- from different parts of the world. They've sequenced 185 Africans from two populations in Africa. [ They've ] sequenced approximately equally [ as ] many people in Europe and in China. And we can begin to say how much variance do we find, how many letters that vary in at least one of those individual sequences. And it's a lot: 38 million variable positions.

But we can then ask: Are there any absolute differences between Africans and non-Africans? Perhaps the biggest difference most of us would imagine existed. And with absolute difference -- and I mean a difference where people inside Africa at a certain position, where all individuals -- 100 percent -- have one letter, and everybody outside Africa has another letter. And the answer to that, among those millions of differences, is that there is not a single such position. This may be surprising. Maybe a single individual is misclassified or so. So we can relax the criterion a bit and say: How many positions do we find where 95 percent of people in Africa have one variant,95 percent another variant, and the number of that is 12.

So this is very surprising. It means that when we look at people and see a person from Africa and a person from Europe or Asia, we can not, for a single position in the genome with 100 percent accuracy, predict what the person would carry. And only for 12 positions can we hope to be 95 percent right. This may be surprising, because we can, of course, look at these people and quite easily say where they or their ancestors came from. So what this means now is that those traits we then look at and so readily see -- facial features, skin color, hair structure -- are not determined by single genes with big effects, but are determined by many different genetic variants that seem to vary in frequency between different parts of the world.

There is another thing with those traits that we so easily observe in each other that I think is worthwhile to consider, and that is that, in a very literal sense, they're really on the surface of our bodies. They are what we just said -- facial features, hair structure, skin color. There are also a number of features that vary between continents like that that have to do with how we metabolize food that we ingest, or that have to do with how our immune systems deal with microbes that try to invade our bodies. But so those are all parts of our bodies where we very directly interact with our environment, in a direct confrontation, if you like. It's easy to imagine how particularly those parts of our bodies were quickly influenced by selection from the environment and shifted frequencies of genes that are involved in them. But if we look on other parts of our bodies where we don't directly interact with the environment -- our kidneys, our livers, our hearts -- there is no way to say, by just looking at these organs, where in the world they would come from.

So there's another interesting thing that comes from this realization that humans have a recent common origin in Africa, and that is that when those humans emerged around 100,000 years ago or so, they were not alone on the planet. There were other forms of humans around, most famously perhaps, Neanderthals -- these robust forms of humans, compared to the left here with a modern human skeleton on the right -- that existed in Western Asia and Europe since several hundreds of thousands of years. So an interesting question is, what happened when we met? What happened to the Neanderthals?

And to begin to answer such questions, my research group -- since over 25 years now -- works on methods to extract DNA from remains of Neanderthals and extinct animals that are tens of thousands of years old. So this involves a lot of technical issues in how you extract the DNA, how you convert it to a form you can sequence. You have to work very carefully to avoid contamination of experiments with DNA from yourself. And this then, in conjunction with these methods that allow very many DNA molecules to be sequenced very rapidly, allowed us last year to present the first version of the Neanderthal genome, so that any one of you can now look on the Internet, on the Neanderthal genome, or at least on the 55 percent of it that we've been able to reconstruct so far. And you can begin to compare it to the genomes of people who live today.

And one question that you may then want to ask is, what happened when we met? Did we mix or not? And the way to ask that question is to look at the Neanderthal that comes from Southern Europe and compare it to genomes of people who live today. So we then look to do this with pairs of individuals, starting with two Africans, looking at the two African genomes, finding places where they differ from each other, and in each case ask: What is a Neanderthal like? Does it match one African or the other African? We would expect there to be no difference, because Neanderthals were never in Africa. They should be equal, have no reason to be closer to one African than another African. And that's indeed the case. Statistically speaking, there is no difference in how often the Neanderthal matches one African or the other. But this is different if we now look at the European individual and an African. Then, significantly more often, does a Neanderthal match the European rather than the African. The same is true if we look at a Chinese individual versus an African, the Neanderthal will match the Chinese individual more often. This may also be surprising because the Neanderthals were never in China.

So the model we've proposed to explain this is that when modern humans came out of Africa sometime after 100,000 years ago, they met Neanderthals. Presumably, they did so first in the Middle East, where there were Neanderthals living. If they then mixed with each other there, then those modern humans that became the ancestors of everyone outside Africa carried with them this Neanderthal component in their genome to the rest of the world. So that today, the people living outside Africa have about two and a half percent of their DNA from Neanderthals.

So having now a Neanderthal genome on hand as a reference point and having the technologies to look at ancient remains and extract the DNA, we can begin to apply them elsewhere in the world. And the first place we've done that is in Southern Siberia in the Altai Mountains at a place called Denisova, a cave site in this mountain here, where archeologists in 2008 found a tiny little piece of bone -- this is a copy of it -- that they realized came from the last phalanx of a little finger of a pinky of a human. And it was well enough preserved so we could determine the DNA from this individual, even to a greater extent than for the Neanderthals actually, and start relating it to the Neanderthal genome and to people today. And we found that this individual shared a common origin for his DNA sequences with Neanderthals around 640,000 years ago. And further back,800,000 years ago is there a common origin with present day humans.

So this individual comes from a population that shares an origin with Neanderthals, but far back and then have a long independent history. We call this group of humans, that we then described for the first time from this tiny, tiny little piece of bone, the Denisovans, after this place where they were first described. So we can then ask for Denisovans the same things as for the Neanderthals: Did they mix with ancestors of present day people? If we ask that question, and compare the Denisovan genome to people around the world, we surprisingly find no evidence of Denisovan DNA in any people living even close to Siberia today. But we do find it in Papua New Guinea and in other islands in Melanesia and the Pacific. So this presumably means that these Denisovans had been more widespread in the past, since we don't think that the ancestors of Melanesians were ever in Siberia.

So from studying these genomes of extinct humans, we're beginning to arrive at a picture of what the world looked like when modern humans started coming out of Africa. In the West, there were Neanderthals; in the East, there were Denisovans -- maybe other forms of humans too that we've not yet described. We don't know quite where the borders between these people were, but we know that in Southern Siberia, there were both Neanderthals and Denisovans at least at some time in the past. Then modern humans emerged somewhere in Africa, came out of Africa, presumably in the Middle East. They meet Neanderthals, mix with them, continue to spread over the world, and somewhere in Southeast Asia, they meet Denisovans and mix with them and continue on out into the Pacific. And then these earlier forms of humans disappear, but they live on a little bit today in some of us -- in that people outside of Africa have two and a half percent of their DNA from Neanderthals, and people in Melanesia actually have an additional five percent approximately from the Denisovans.

Does this then mean that there is after all some absolute difference between people outside Africa and inside Africa in that people outside Africa have this old component in their genome from these extinct forms of humans, whereas Africans do not? Well I don't think that is the case. Presumably, modern humans emerged somewhere in Africa. They spread across Africa also, of course, and there were older, earlier forms of humans there. And since we mixed elsewhere, I'm pretty sure that one day, when we will perhaps have a genome of also these earlier forms in Africa, we will find that they have also mixed with early modern humans in Africa.

So to sum up, what have we learned from studying genomes of present day humans and extinct humans? We learn perhaps many things, but one thing that I find sort of important to mention is that I think the lesson is that we have always mixed. We mixed with these earlier forms of humans, wherever we met them, and we mixed with each other ever since.

Thank you for your attention.

(Applause)