so in the talk today I'm going to tell you about evidence that's really emerged very strongly over the last eight years really even the last four years that shows that human population mixture repeated large-scale mixtures are very different and not so different human populations again and again over time is an important theme of our past it's not just occurring in the last five hundred years when we have many documented cases for example in the Americas but it's occurred thousands of years ago in tens of thousands of years ago and hundreds of thousands of years ago it's a central theme of our past so I'm going to begin by introducing a bet by luca cavalli-sforza who died a couple of weeks ago who was the pioneer a field of genetic studies of the human pass cavalli-sforza was a really extraordinary creative Italian population geneticist who realized that the differences amongst human populations around the world today could be used to reconstruct the history relating those populations by studying the genetic differences that exist between individuals and on average across populations that must reflect events that have occurred since those populations separated from each other and by studying those patterns he could determine with his colleagues which populations are most closely related to each other genetically and reconstruct trees of population relationships and his bet when he started out this work in 1960 when the genetic technology was very primitive in fact there was really no genetic technology for studying variation instead you might be looking at variation in blood group types like the polymorphism that you get typed if you're studying for organ transplantation for example was that the populations distributed around the world today with the exception of the great movements known in the last 500 or a thousand years reflects the people who dissent from people who got there were a very long time ago and by studying the differences amongst populations in the world we can reconstruct how people got to each place in the world over time and so he made this bet that would be possible to reconstruct the past based on the history of present-day people and over the subsequent decades one study after the other gather data attempting to realize this idea so in 1978 and here's an updated analysis of this type published in 1993 he analyzed blood group type variation like the blood group in Europe and studied the main gradients of genetic variation the frequency differences amongst populations in the data set at these blood group types and what he found was that the single biggest predictor of differences in blood group types was a gradient moving from the southeast of Europe near Turkey all the way to the northwest near these islands and he argued that what is seen here and what the genetic patterns are reflecting is the history of farming migration into Europe we know from the archaeology that farming which was first invented than they are East twelve to eleven thousand years ago moved into Europe around nine or eight and a half thousand years ago and proceeded to spread across the continent the idea was that these farmers from the Near East would have mixed with the local hunter-gatherers who were always already established in Europe and their ancestry from the Near East would be gradually diluted as they spread west and north across Europe until today we would have much lower percentage as a farmer ancestry in the northwest much higher in the southeast but now from ancient DNA where we can go back and obtain whole DNA sequences from people who lived at the times in question we now know that the gradient of farmer ancestry is almost exactly perpendicular to what was argued based on the interpretations today that is that main gradient of variation is not driven by the farming expansion but by other things and the reason for this is that there were very important population replacements and migrations that happened after the spread of farming I'm going to tell you that about that in maybe 20 minutes what the evidence is for that so what's actually become clear is that even though it was very exciting idea that you could reconstruct the past based on patterns of variation in people living today in fact it's impossible and the reason it's been possible is that because people have mixed too much they've moved too much again and again and it's overwritten the populations that existed there before and so the only way to get back to what those movements were like is to actually sample the people who lived at each time and study how they related to each other so what happened really beginning in 2010 was a new scientific instrument got invented or became practical so the technology for ancient DNA has been around for several decades but really in 2010 it became possible for the first time to obtain really meaningful ancient DNA data on the scale of an entire genome and to regularly extract DNA from ancient skeletons of quality similar to people living today and so when a new scientific instrument is invented like a microscope or a telescope for example in the 17th century when the microscope was invented it makes it possible to peer into worlds that were never previously known for example when the microscope was invented to look into cells for the first time to look at microorganisms and to see a world that was not even realized to be there before and in the same way ancient DNA by looking at people in the past from ancient peoples who have never been interrogated before in terms of how they relate to each other and to people living today they find surprise after surprise and one guess after another about how ancient people related to each other turns out not to be true so the measure of how revolutionary a new scientific instrument is is how often when you look at data for the first time that's not been looked at with this instrument it reveals something that's not known and it's very clear that this is a revolutionary instrument in that sense and I'll show you why so this new technology of ancient DNA begins with a bone that we take into a clean lab where we try to protect the bones from contamination by the technicians and the other people who are handling the bones because we have much more DNA than is left in these phones and contamination from us will confuse our analyses we drilled beneath the surface of the bone again to try to remove contamination that would be introduced to the bone through people handling it we obtained some powder some tens of milligrams of powder we release the path the DNA from the powder in a watery mix that removes the protein and the mineral content and then we sequence it in one of these modern sequencers that first became available in the late 2000s and by now have reduced the cost of sequencing DNA on a per letter basis by literally about a million fold since 2000 and since the year 2000 and this makes it possible to ask and answer questions about the past that were simply not addressable before so this is just a picture of illustration of Moore's law which has nothing to do with genetics but is in computer science which has observed that the increased density of elements on computer circuits of transistors on computer circuits has doubled every one to two years for approximately the last century and this has been central to driving the computer revolution in the same way there's been a Moore's law of ancient DNA but it's been much faster so there was the first genomes were published in 2010 it reached ten individuals with whole genome data in 2012 it reached more than a hundred in 2015 it reached more than a thousand just this year in our laboratory that's the slides out of date there are now more than 6000 working genome samples and so it's increasing very rapidly in our laboratory and other Radlett laboratories and this order of magnitude increase in whole genome sequences providing data about the past makes it possible to address problems that were simply not addressable before so I'm going to give you one slide that tells you a little bit about how genetic sequence data might provide information about the past of course this is a quick summary so what we're looking at is DNA which is obtained from almost all of our cells not quite all of them carry our genome which is packed into three pairs of chromosomes that all together in length are about three times three billion DNA letters long since you have two copies one from your mother and one from your father about six billion DNA letters long and they are strings of four letters adenine cytosine thymine and guanine AC T and E 99.9% of the DNA sequence you get from your mother and the one you get from your father are identical but in that point 1% there are differences in over a three billion long base letter long genome that's about a three million differences and those differences reflect mutations that have occurred in the past they've built up randomly over time as shown here between the genome sequence you have of your mother and your father and kind of tell you how long it's been since your mother and father's DNA share a common ancestor because we know approximately how long it takes for mutations to accumulate if you see a dense run of mutations that might tell you that your mother and father's DNA are related very far back in time about a million years and if you see not many differences it might say they're only related a few dozen generations ago and so by comparing pairs of genome sequences to each other and looking at the density of mutations as it varies across the genome we can reconstruct variability of across our ancestors and how we're related to each other so the genome is not just a single person from a statistical point of view you're not a single person you contain a multitude of ancestors so it's this is very confusing often but in fact when you look at your DNA you think you're a single sample but in fact you descend from two parents and your parents descend from four grandparents a you descend from four grandparents eight great-grandparents and so on going back in time and if you go back even ten generations you have about a thousand ancestors if you go back 20 generations you have a million ancestors and when your mother and father produce a sperm or an egg they split the DNA they got from their own parents one or two times per chromosome per generation and so the DNA that's coming to you is fragmented across those ancestors and so when we sequence our genome were interrogating thousands tens of thousands maybe up to a hundred thousand ancestors and so you're not a single person you contain a multitude of ancestors and as a result by sequencing a single person's genome you can position that person's DNA with exquisite accuracy relative to other genomes that have been sequenced it's not like flipping a coin once in order to figure out the probability that it's a heads it's like flipping a coin a hundred thousand times and then you can get a very accurate sense of the probability that that coin has heads 50 percent perhaps and so that is why there is so much power in a genome because it's not studying a single person it's studying and and and interrogating and sampling a multitude of ancestors so in my talk I'm gonna tell you about four examples of how we were wrong and in particular I was wrong in my understanding of the world and how the genetic data that I had the privilege to look at surprised me and my colleagues very strongly and forced our hand to think in new ways so I'm going to begin by talking about the findings related to the sequencing of archaic humans work that was led by Svante pay bows lab in Germany and work that I was intensely involved in as a person analyzing the data since the data began to be generated in 2007 so when I was a PhD student I was trained in a laboratory that was in fact a student of luca cavalli-sforza that was part of the triumphant establishment that modern humans all around the world today Africans non Africans are descended from a common ancestral population that lived in Africa 50 or 100 thousand years ago or more and that all of our ancestry into the best we could tell at that time seems to descend from an African source population now that's still true that the great majority of our ancestry of everybody living today derives from African lineages deep in time but at the time it was almost seemed like an absolute rule and I was very much in that school when I entered this work so modern humans anatomically modern humans people whose skeletons look like ours first for the appear for the first time 200 to 300 thousand years ago in the fossil record of Africa and don't really appear outside Africa in the near East until after about fifty thousand years ago so there's this explosive spread of modern humans Out of Africa in the near East after fifty thousand years ago into places like Europe where this is a skull of your early European modern human um Neanderthals meanwhile were archaic humans that first appear in the skeletal record beginning at least about four hundred thousand years ago and persisted in Europe at least until forty thousand years ago and probably disappeared shortly after that Neanderthals had brains that were as large as ours as measured by the size of their skulls they had bodies as big as ours and they made tools as complex as our own ancestors did they were clearly very sophisticated people but they disappeared after modern humans and expanded into their territory after a modest period but not trivial period of overlap when these very different people who we now know separated from each other half a million years ago or seven hundred thousand years ago encountered each other so a question that's always been in the air ever since it was shown that Neanderthals and modern humans for example in France for example dour caves where there's layers of Neanderthals and then modern humans and Neanderthals they were surely encountering each other when they encountered each other over this period of time was there interbreeding between these two very different groups of people more different from each other than anybody in the world today but both intelligent and sophisticated groups of people and if there wasn't of reading where did it occur did it occur in the near East which was the first place that there's documented overlap or in Europe which is the place where the record of Neanderthals is more dense or is there mix blur with archaic humans and modern humans in eastern and southern Asia where there is less clear archaeology and where in many places there's no evidence of Neanderthals at all but maybe there are other humans that certainly left tools and perhaps encountered modern humans so when I got involved in this Neanderthal study sponte people and colleagues were generating whole DNA sequences from Neanderthal and we and our colleagues applied the world's simplest test to try to understand how the neanderthal genome we were generating is related to modern humans and so this picture is a cartoon of one of our 23 chromosomes so that lobe that elliptical lobe on top is one arm of our chromosome the elliptical say your chromosome 3 the elliptical lobe on the bottom is another arm of your chromosome 3 that circle in the middle is the centromere it's a structural element that holds the that's involved in pairing chromosomes together and that would be packed with hundreds of millions of DNA letters of DNA letters so what we do is we compare the Neanderthal genome to two modern human genomes a French person in this case in a Nigerian person and we look at positions where the French and Nigerian person are different for example one has a DNA letter thymine T and another has a DNA DNA DNA letter guanine G and then we lift the veil on the Neanderthal sequence that we've generated and we ask does it have a T or a G and it should match equally often if modern human ancestors separated from the ancestors of Neanderthals and there was no subsequent interbreeding and when we look at it there's ninety two thousand matches or so to the French person and eighty four thousand matches or so to the Nigerian person and this is a very statistically significant excess it could not occur by chance it's not just french and Nigerians it's all non Africans and all sub-saharan Africans and so we looked at this many many different ways not just this way but other ways and the only possible explanation for this was interbreeding of Neanderthals into the ancestors of non Africans but not as much or at all into the ancestors of sub-saharan Africans we also were able to obtain an estimate of how much of the aunt of the DNA of non-africans today it derives from the Ender Falls and the way we did this is we carried out the same analysis where we saw remember there was ninety two thousand two eighty four thousand so there was an excess matching to Neanderthal in there was an excess matching to neanderthal of non Africans and then we replaced the non Africans in that analysis with a second Neanderthal we got data from and we saw how much of the way to a full Neanderthal is a non African person today and it was 2% of the way and so in that way we can obtain a formal and statistically valid way of seeing what percent of the genome of a non African person is a is Neanderthal it's a little bit higher and significantly higher in East Asian people than in European people and there is concrete and very interesting reasons for that so the next thing we were really interested in doing is understanding when this mixture between the Andrew fells and modern humans occurred and what we did here is we took advantage of the phenomenon I told you about before which is that when you produce a sperm or an egg you chop up your chromosomes that you got from your mother and father so here again are as a cartoon of your chromosomes so imagine this is your chromosome free I show two lines of red to show that you carry from both your mother and your father to modern human chromosomes and two lines of green to show that your partner carries their mother and father from a Neanderthal then the kid has one entirely Neanderthal derived chromosome one entirely modern human derived chromosome and then that kid breaks up their chromosome one or two times per generation as they make a sperm in egg and pops it up and we know the rate at which it shots gets chopped up and by looking at the dice size the typical chunk size of Neanderthal ancestry we can read off how long it's been since modern humans and Neanderthals since this Neanderthal type ancestor was introduced into modern humans and so if you look in a Nigerian person and this is a curve of how long these segments extend that's a very slow decay very fast decay there's very little segments and these reflect very old shared ancestry before the the groups diverged from a common ancestor for the most part but if you look at non Africans today you see a very much longer fall-off and that reflects the fact that there's big chunks of Neanderthal ancestry and French people which are many times more recent than what you see in Nigerian people if you then look at a forty five thousand-year-old modern human genome from Siberia that we in our colleague sequenced the chunks are huge that's the upper curve there compared to modern Europeans and that's because that individual witnessed or was very close to the time of mixture was then maybe 7,000 years of the time of mixture so by using this type of data combined with the direct dates on these ancient skeletons like this 45,000 year old individual we know actually rather precisely about when the next year between the and refills and modern humans occurred and it occurred at the dawn of the explosion of modern humans Out of Africa between rather tightly fifty four to forty nine thousand years ago so the conclusions of this Neanderthal interbreeding section is that interbreeding happened we think that there was an interpreting event in the near East where we think there is at least circumstantial evidence that Neanderthals and modern humans overlapped it's left a legacy of about two percent Neanderthal DNA ultimately in that population and spread out all around the world and is the primary source population of non-africans today in everywhere we're not Africans today live except for people who are sub-saharan African ancestry who have moved there since that time so what you encountered also in 2010 an amazing surprise so while we were finishing up the neanderthal genome project Svante Paabo and his colleagues received a bone that was marked modern human from colleagues in russia from this site in the Altai mountains in South Central Siberia it was a finger bone and this sample had 70 percent of the sequences were primate human type DNA and when we and their colleagues there was another tooth that also had DNA of this type when we sequenced its female derived DNA it was related to when we started when we sequenced its whole genome we found that it was more closely related to Neanderthals than to modern humans but very distant from the rest of the Neanderthals and because of this it was an entirely new archaic population so we took this new archaic clean own which we had stumbled on completely at by accident this is an example of the destructive power of DNA with Neanderthals we knew they were interesting it was of course exciting to obtain the sequence of them and compare them to modern human we had a very well posed question did interbreeding occurred but there wasn't even a question for this other group there was no skeletal record there was no well-defined archeology so in the case of Neanderthals it would a fossil in search of a genome we knew it was interesting to attend a genome but in the case of these denisa bins we stumbled on the genome and we still don't have a fossil record associated with them all we have is this little finger bone and a few teeth and now a bit of a skull cap and we don't know about their morphology the shape of their skeleton we don't know what type of tools they made except we know they were close cousins of us and of Neanderthals so the first thing we did with the data was we applied some of these same ideas that we had developed to apply to Neanderthals and we asked to they match some humans more than others and one of the most interesting findings of this analysis was when we did the following when we took a New Guinean genome from Papua New Guinea which is north of Australia and compared it to a Chinese genome and looked at differences and we asked whether Denis events matched the New Guinean or the Chinese more there was a very significant access mapping to New Guineans and the only way to explain this is gene flow from people related to these Siberians into the ancestors of New Guineans more than into the ancestors of Chinese and the proportion was about 5 percent 4 percent of the ancestry of people in New Guinea today derives from an archaic group related to these Denyce events so subsequent work we did has looked at the proportion of Denyce of an ancestry in different parts of the world and if a full shaded circle is 5 percent what you see in New Guinea and the indigenous people from the Philippines in Australia there's a small amount in East Asia and almost none in Europe and in Africa so it's very clear that modern humans as they moved east encountered these Denise 'van populations and that there's a large proportion or a substantial proportion affecting groups like New Guineans and some and we now know an independent partially independent event in Northeast Asia and this is a second and unanticipated or even a two additional unanticipated mixtures with archaic humans that we didn't know about the conclusions of this is that there were two interbreeding events at least and we now know there were several additional ones including related populations one of these Neanderthals who have this incredibly well-documented archaeological record that we've known about for 150 years and elaborated since that time as a scholarly community and the other these Denice events which we have no idea about from the fossil record and it still has not yet been done an analysis where this genome type is connected to a skeleton which has a morphology and a type of tool use that we can assign to it so that's something to prompt archaeological and excavation work in order to connect these two just published a few weeks ago an amazing thing from Svante Paabo and colleagues not work that I was involved but the genome was sequenced again from this site and aeneas of a cave of a girl whose mother was a Neanderthal read here indicates Neanderthal but her father is a Denis event so this is a first-generation mixture of these two groups separated from each other for 400,000 years but they were coming together at this cave we actually have a number of humans from 30 40 50 60 70 80 thousand years ago who are mixtures within one to six generations of different very divergent humans Neanderthals Denisovans and modern humans and the fact that we're seeing such a substantial fraction of these genomes that were analyzing being mixtures or close to the time of mixture suggests that maybe mixture between these very different groups was common in our past that when they encountered each other people did tend to mix and frequently so what's happened is that the DNA has unleashed perhaps what you might think of as a Pandora's box of mixtures we didn't know about without DNA between archaic and modern humans and these are some of the known events moving between mixing the cross leaves very different lineages now documented by DNA and additional ones are being published every few months or year these days so the second part of my talk is I'm going to talk about evidence that no one is pure and I'm going to talk particularly about the part of the world were in right now which is Western Eurasia Europe is part of Western Eurasia the West the Near East is another part of yet Western Eurasia and I'm going to talk to you about the results of a study published in 2016 which analyzed hundreds of whole DNA sequence says from people of West Eurasian ancestry and ask the question how does the DNA of those people at that time in the past that we have data from relate to people in this region today so just to give you a like a very quick overview of genetic variation in Eurasia today if you look at frequencies of variable positions in the genome where we differ from each other there's a vast region of relative homogeneity extending all the way from Iran and Central Asia all the way to the Atlantic facade of Europe here for example where people are genetically pretty similar to each other and then there's a transition zone and people of region of East Asians where you have a similar phenomenon of a broad region where there's over a vast region people who are genetically pretty similar to each other although there's important differences and in South Asia yet another broad zone of quite a lot of difference but homogeneity within and so this pattern for hasn't been noticed for centuries just even a based on how people look which is not a very reliable thing but it's prompted the idea of the word quote Caucasian to refer to this group of Western Asian ancestry but what we now know from genetics is that this group just didn't exist ten thousand years ago and I'll show you the evidence for that so we got DNA from hundreds of of individuals who lived ten thousand six thousand three thousand years ago and compared them to people living today and if you measure at the top level of this picture the average differentiation in terms of the frequencies of people of genetically variable positions and people living today what you see is that the level of differentiation is about 0.098 or 0.1 on this scale what you're measuring is basically you're looking at the frequency of a cytosine or as guanine a DNA base in each population and you're taking the square that that frequency difference between say population a and population B and you're averaging it across hundreds of thousands of independent positions in the genome that you're looking at and if you look at that measure which measures how similar populations are to each other you get that ten thousand years of Crowe across western Eurasia the typical level of differentiation was around 0.1 and that's the same as Europeans a nice day today so across western Eurasia today the level of differentiation is about 0.01 that's the bottom of this plot and at that time it was that's different from Europeans and East Asians by six thousand years ago it had dropped by a factor of three and by about four thousand years ago it had dropped by a factor another factor of three and it's not because populations went extinct or disappeared we now know what happened so in this paper we also showed that almost all the DNA not quite all of it but almost all the DNA of people in West Eurasia today comes at the time depth of around ten thousand years from for highly different sources each of them is different from each other as Europeans and East Asians and each of them we have DNA for one of them you could think of as Iranian earlier of early farmers or agriculturalists who lived about nine or ten thousand years ago one of them you could think about is Levantine farmers or early agriculture list people who lived in Jordan and Israel for example and around that time one of them is Western hunter-gatherers people who are like cheddar man or people who lived in this part of Europe about that time and another is Eastern hunter-gatherer of people who have lived in the far east of Europe in these groups none of them disappeared instead they mixed with each other and that's what you see today and the reason for this homo homogeneity is a massive mixing since that time where all these groups contributed to people living today so lesson 3 in how we were wrong relates to a findings about how large-scale migration was common in our past and I'm again gonna talk about Europe because Europe is a place where we happen to have really amazing quality data not because Europe is more important than other parts of the world it's not it's only a small part of the world but it's a place where there's good archaeology and where ancient DNA studies have happened to be a little ahead of where they are elsewhere in the world and so we get a flavor for what it can show about the past so as I mentioned earlier in the talk agriculture farming it invented for the first time in the world in the near East twelve to eleven thousand years ago shortly afterward independently probably in East Asia but West Eurasian agriculture from the Near East gets to Europe a couple of thousand years later around eighty five hundred years ago jumping probably from Turkey anatolia to greece and then spreading out so it reaches within two and a half thousand years these islands and also Scandinavia so there's another cultural phenomenon not documented by archaeology but documented by the languages people speak and in fact people in Europe with a few very interesting exceptions speak extremely similar languages so English is an example of an indo-european language which is related closely at some level to the Latin and Romance languages to Germanic languages to Greek and even to languages in India and Iran which I will talk about later and so a question is how does this language homogeneity come to be in other parts of the world for example there was small region of the Caucasus in South between the Black and Caspian Seas there's actually much more language diversity within that region and then across the entirety of Europe in much of the South America there's much more diversity and much deeper diversity than in Europe so it suggests that some process happened to spread relatively homogeneous languages that over not so long a time diverged from each other and became a little bit mutually in unintelligible as they are today so it was thought at the over the primary hypothesis prior to 2015 that was dominant was that it was the spread of agriculture and the movements of people that this new economic system would have allowed from the Near East that probably spread these languages into Europe and we now know that's wrong because we now know there was another large scale migration and they'll tell you the evidence for that so ancient DNA a few years ago had initially been available only for hunter gatherers and farmers so people who lived before about four and a half or five thousand years ago in Europe had only two types of ancestry ancestry like those of the early Anatolian farmers which actually we didn't yet have data from at that time but first farmers of Europe or first hunter-gatherers of Europe but if you look today there was a third component of ancestry shown in green in these bar plots showing breakdown of sester types in each part of the world where for example in northern europe about half the ancestry was of this green type which simply didn't exist in europe 5,000 years ago so how did it get there so in 2012 we and colleagues we in our laboratory we were we had we developed a new statistical test to see whether a population was mixed and the way this test works was that we took hundreds of thousands of positions in our DNA where people vary and we looked at the frequencies in each population in the world so for example in French people we would measure the frequency in French people of a cytosine at a particular position it might be thirty percent and then we might compare it to Sardinians people in southern Europe and to Native Americans and if on average the French are intermediate between Native Americans on the one hand and start Indians on the other hand that means for sure that the French people are mixture of two groups related maybe very distantly in the past to Sardinia the Native Americans so we took data from 50 populations around the world and we applied this test to each of the 50 populations and we took all other possible pairs of those 50 populations and we saw are they good candidates for a mixture for each of them so when we did this to French and we looked at all possible pairs of populations we got a huge and absolutely convincing signal of mixture between Native Americans and Sardinians and this is we of course didn't interpret this it was certainly that was the correct answer it wasn't Siberians and Sardinians and it wasn't South Asians and Siberians it was very much Native Americans and what we didn't we didn't think that what was going on was actually that's Native Americans crossed the Atlantic and mixed into Europe instead what we proposed was the following we proposed that there used to be a population that we called ancient North Eurasians a population that was once distributed across northern Eurasia and event including Siberia some time before fifteen thousand years ago it moved east and contributed to the population that people be Americas and sometime later it also contributed to the people that peopled Northern Europe and contributed to French people today more than two Sardinians explaining our signal and so this is what we proposed and that Northern Europeans today are a mixture of Sardinian like people who we might think of as descended from the first Farmers of Europe and we've now documented that that's probably a good approximation and an ancient North Eurasian population and so we predicted this ancient North Korean what we call a ghost population it's a population that we predict statistically based on patterns it's left in people today but that doesn't exist in unmixed form anymore and then the subsequent year SK will err sleves group working in denmark dot DNA from a 24,000 year old boy who lived at that site malta in south central siberia and found that that individual was closely related to native americans and to europeans but not to people who live in that region today showing that they're a representative that of that ancient north eurasian population they had actually gotten DNA from this population and that was an even better proxy for this predicted population then then the Native Americans and so we now know that somehow this ancient North Eurasian ancestry entered Europe and we wanted to find out where so what we did is we collected DNA across this whole time period so here is the genetic variation of diverse present-day West Eurasian population so each dot is a person there are about 800 dots and what we do in this analysis is we take a grid of data so we take about 600,000 positions where we have looked at the DNA we take about 800 people so it's a matrix of eight 600,000 rows and 800 columns and we multiply the matrix by itself to reveal how close each individual is to each other on average and then we look at the main gradients of variation much like cavalli-sforza did in that first plot I showed you and what you see is a really remarkable pattern where there's two parallel gradients with relatively little in between on one side of the plot on the left are Europeans and the right side of the plot are near Easterners on the top are people with more ancient north Eurasian ancestry and on the bottom there are people with less and we now know how this came to be and I'm going to gray out those dots so here's these two parallel Radiance and I'm going to show you the ancient people of West Eurasia over time so here are the hunter-gatherers the hunter-gatherers fall beyond Europe in the direction of Europe in differentiation from the Near East and what that it reflects is the fact that Europeans today are a mixture of these hunter-gatherers who no longer exist in on mixed form and ancient near-eastern earth and that's how Europe forms in between these two groups the next thing that happens is you have a large number of farmers from first Farmers of Europe and they fall genetically on top of where Sardinians fall today we're consistent with the ideas Sardinians are isolated descendents more or less of the first farmers of Europe but you still don't see people like in the main part of that the gray dots on the left who are like Europeans today meanwhile in Far Eastern Europe on the steppes north of the Black and Caspian Sea this other population develops as a mixture of Europe Eastern European hunter-gatherers and Iranian hunter-gatherers who are at the top end of the Near Eastern gradient but you still don't see people like most Europeans today and it's only after about 5000 years ago when you for the first time see DNA of that type and what that's reflecting is movement from Eastern Europe of these people who are green squares at the top into Central Europe and then subsequently for after that you see populations like you see today we now have hundreds and hundreds and even thousands of individuals that are very much like people in Europe today so in summary there is evidence now from genetics li-like to mass migrations the first after 9,000 years ago into Europe bringing farmer ancestry shown in orange here there's some mixture with a local hunter-gatherers shown in blue and then after five thousand years ago there's a mass movement of people who are from steppe pastoralist backgrounds and they form north of the Black and Caspian Sea about five thousand years ago and they spread through Europe so the people who spread this ancestry and originated this ancestry are an amazing group archaeologically it's a group called the yam nya so the yam nya are the first people who use a new economy and the open steppe lands the we all had shortly before had been invented and the horse that shortly been domesticated and they hitched horses to wagons wheeled wagons and use them to bring supplies and water out into the open steppe and graze their animals on the open grasslands that were available and much more intensively exploit the step before this the steppe was a patchwork of very different archaeological cultures that made different artifacts in different parts of this vast region after the omneya phenomenon many of these groups disappear and are replaced by a homogeneous culture of people making pots and artifacts in the same way and leaving very few villages because these people were mobile pastoralists but leaving very rich graves in which these people were buried these people spread over a vast territory all the way from Mongolia in the East to Hungary in the West and the genetics shows that they also spread into Europe and are the single primary most important contributor to Europeans today so an implication is that indo-european languages almost certainly came with the second migration and there's now evidence as I'll tell you at the end of the talk that this type of ancestry from these young Naya steppe pastoralist eventually got into South Asia today so a plausible hypothesis is that these people are responsible for spreading these spoken indo-european languages sometime after 5000 years ago so in Britain this ancestry reached a couple of hundred years later and so we published a paper earlier this year where we analyzed data from people who made these pots so these are these bell beaker pots that were first appear in the archaeological record probably in Spain about 4700 years ago and then by 4,500 years ago are all over Western and Central Europe and in Britain very heavily and in Ireland and France and southern and in northern Italy in Hungary and Switzerland even as far east as Poland and is associated not just with pots but many other types of grave artifacts and ways of burying people archaeologists off interpret this as a as a religion or a belief system that's reflected in a particular way of bearing people is a very dramatic Western European expansion from west to east so we obtained data for more than 200 individuals who were buried with these pots in their graves many of them from Britain many of them from elsewhere and we were interested in what the spread of this culture how this related to the spread of people and when we looked at the genetic data in the same type of analysis we found that different from the omnia this group I told you about before these people were not homogeneous this culture spread by acculturation by copying by people spreading ideas not entirely by people in particular if you look in Iberia Spain the genetic ancestry of people who buried with these pots is almost entirely different than that in Central Europe so somehow this religion or belief system spread between these two regions by copying kind of like cell phones spread between people of different ancestries but once it got to Central Europe it's spread by movement and I'm going to tell you what happened here on these islands so in around 4,400 years ago just after the last stones at Stonehenge got big stones at Stonehenge go up these this these these bell beakers and the associated artifacts begin to appear in Britain and here's an example of a bell beaker burial and what you actually see is that going from the top to the bottom of that plot blue shows Farmar ancestry suddenly at 20 40 400 years ago there's an almost complete population replacement so we have dozens and dozens of samples before this time and dozens and dozens and dozens of samples after this time and in red you suddenly see this step ancestry for the first time and we can't find a single burial that has entirely former ancestry anymore it's a ninety percent population replacement it's the it's a permanent replacement this is still the primary ancestry of people who lived most of the people who live in Britain today so an even more amazing thing happen in Spain this is unpublished data but if you look at the proportion of steppe pastoralist young Naya related ancestry that's on the y-axis here in people in Spain prior to twenty to forty five hundred years ago 2500 BC it's all constant it's just farmer type answer three but beginning about 2500 BCE there's also people with steppe ancestry coming in and the coloring of these dots is only done on the males and it's their Y chromosome it's the DNA you get from your father and he gets from his father and so on and what you see is that these people with elevated step man's history always have step-like chromosomes and then these populations mix together you see after four 2000 BC four thousand years ago and you get a population which we can accurately determine has about 40% ancestry from these people with steppe ancestry and 60% ancestry from the local Iberian farmers so it's a forty percent population replacement or movement it's a lot but it's not most of it but there's a complete Y chromosome replacement so that means males coming in had preferential access to local females again and again and again and over the generations so it's telling you something about culture it's telling you something that the collision of these two populations was not a friendly one not an equal one but one where the nails from outside were displacing local females and did so almost completely so the fourth and final thing I'll talk about is the history of South Asia of India and Pakistan and neighboring countries in 2009 my colleagues and I analyzed data from modern South Asians diverse people and published a notice that the great majority of people in India today can be described with that with important exceptions can be described as a mixture of two highly divergent populations each of them is different from each other as Europeans in East Asians are today and that people in India today in different groups that are often endogenous have different proportions of ancestry from these two ancestral populations ranging from about 20% of the one that's more closely related to Wester Asians Europeans Central Asians and their Easterners to 80% of that ancestry in that proportion of ancestry we're not we showed was related to the language people speak so most people in India speak into European languages which are concentrated in the north but hundreds of millions of people speak for vidiian languages in the south so indo-european languages tend to be spoken associated with more of this and North Indian West Eurasian related ancestry and within any region people who are of traditionally higher social status in the traditional caste system have significantly more on average with important exceptions ancestry related to West Eurasian so this is clearly telling you that the collision of these populations with these different ancestry types what we call ancestral North Indian and ancestral South Indian was a cultural phenomena phenomenon where again there there are differences of culture that are left in the caste system today for example in languages that were associated with this mixing process so last year we published a much bigger dataset with about two hundred and fifty groups analyzed that hundreds of thousands of positions in the DNA from all over India which is an incredibly diverse place in terms of human variation and we used it in now in telling you this partially published work to test our model of Indians and our model initially in 2009 was as follows that people in India today are a mixture of two very different population one of which is West Eurasian related and the other of which would had no ancestry from people outside the Indian subcontinent and had no West Eurasian related ancestry at all was more closely related to East Asians or to indigenous people of the Andaman Islands and Australia and New Guinea and so when we analyzed the data we looked at the possibility that we looked at the present day Indian groups which are shown in little plus signs here and we now know through further analysis that actually these two source populations for India at these two poles at the top and the bottom of the line in fact are themselves composed of three populations one of them is Iranian farmers like people another step pastoralist type people and the other is Asian people not particularly closely related to the continent but if you look at this line and intersect where it is if you look at people who have no steppe ancestry at all they actually clearly have about a third of their ancestry or a quarter of their ancestry being Iranian type so one of these ancestral groups was do not have knowest West Eurasian ancestry at all it had a big chunk of ancestry from people related to Iranian to farmers and similarly if you look at people with no Asian ancestry at all they were a mix of Iranian related people and step-like people and so we used DNA and so the ancestral South Indians would be at that extreme and the ancestral North Indians would be at that extreme and in fact there are people in India today who are almost unex descendants of these ancestral South Indian populations so to learn more about how this form we've gathered and still haven't published data from five hundred and fifty ancient genomes from Central and South Asia from Kazakhstan and from Uzbekistan Turkmenistan and from Pakistan and Iran and there's about two hundred and individuals from Kazakhstan and Russia from the forest and steppe zone is another two hundred or so individuals from Iran whose Pakistan Turkmenistan and so on most of these samples spanning between about 7,000 years ago and 3,000 years ago and then another 140 samples almost all from a small region of northern Pakistan the Swat Valley so one of the most interesting sets of samples are these samples from the bacteria margiana archaeological complex so this was an ancient civilization of the ancient world that was spread between about around 4000 years ago in Turkmenistan and Uzbekistan north of South Asia and was really only discovered in the 1990s that it existed and maybes huge cities or towns like this one at gona or Taipei and we got data from four different sites and these individuals are genetically extremely similar to each other buried at the cemeteries from these sites most individuals are genetically similar to Iranian 's with some important mixture events that we use in our analysis but there are a small number of outlier individuals who have very different ancestry which reflect people they were contacting before four thousand years ago they looked like local hunter-gatherers from Siberia that we also have data from and from Kazakhstan but after four thousand years ago they look like people of the steppe at that time derived ultimately from the yayaya so we can actually see the impact of the steppe pastoralists descendants coming through this region after four thousand years ago and then in Pakistan we see evidence of them having already mixed in by thirty five hundred years ago so we've now in in South Asia as in Europe documented a chain of transmission of this step pastoralist ancestry from the yam nya through a series of intervening steps into Europe and into South Asia so one of the most interesting things that we have is from two of the sites we analyzed one of them was this go Norte which I showed you and the other was a site called char Ashok down eastern Iran we now have data from 14 individuals who are outliers from the main clusters at these sites and around 4,000 or older than 4,000 years ago and these individuals have no steppe ancestry and they don't even have the type of Iranian ancestry at the site it's clearly a very different type of ancestry and instead they have mixture from South Asian hunter-gatherer type ancestry with a different type of Iranian ancestry it's a very distinctive ancestry type we call this the Indus periphery population because these are sites in cultural contact with the Indus Valley Civilisation which was contemporary in to the south of the beating of the bacteria marciano carp archaeological complex it was one of the first great civilizations of the ancient world and we think we're actually seeing immigrants from that civilization we are characterizing the ancestry it was a cosmopolitan population people have different proportions of South Asian indigenous hunter-gatherer type ancestry and Iranian type ancestry and we think we've defined a gradient of that ancestry so going back to this triangle we now have three proximate sources for the ancestry of Indian people today and we can use it to reconstruct a history over time so the first is this industry population where we have samples like those red dots from these outlier sites that define a gradient of ancestry between South Asians hunter-gatherers on the one hand and Iranian farmer type people on the other hand and it's established defines a gradient that probably existed in fact we've documented existed of more than 4,000 years ago in this region and that was the primary gradient of ancestry we think in this region once the steppe pastoralists come in after 4000 years ago a new gradient is established and we've documented that with these Pakistani samples where people are mixing with the steppe pastoralists in the bottom left of the and one of these stamped one of these populations along this gradient but you still haven't formed the main gradient in India today and that forms later so there's three gradients that form over time and we're only looking at the most recent level of those gradients just to summarize there's actually these what's interesting to think about this but actually what I've told you about is a tale of two parallel protrusions from Eurasia subcontinent's one being Europe and the other being India they're similar size similar populations long term over time may actually have remarkably parallel histories from a big-picture way agriculture expands in the near East about eleven thousand years ago spreads into each of these regions sometime after 9,000 years ago in South Asia from the East from Iran in Anatolia in Europe from the West from Anatolia over the next few thousand years this type of ancestor this technology spreads across each of these regions and we've documented with genetics that it probably involved mixing in both region meanwhile these young Naya steppe pastoralists after 5,000 years ago spread into both regions with slightly different timings in each place creating another gradient and mixtures of these mixed populations in both places form the primary driver of variation in both regions today so a summary is that ancient DNA is teaching us that much of what we thought about the past is wrong it's almost always true that when we look at DNA from a archaeological time period and culture that we haven't looked at before it's very surprising and different from what our orthodoxies were before about the past that I think the it's likely to be through that we're all mixed at many levels far back in the past and none of us are pure and in fact mixing in migration or in our nature and I think we should think about that in these times and this is an unusual field compared to many other areas of science we're scratching the surface is guaranteed to surprise and particle physics for example when you study you have to build a bitter Bigler and bigger accelerator and many of the things have been figured out and with much much work you might be able to answer a specific question and somehow you know I always thought I was loved science but I always wanted to get into this air yeah it's a science but it always seemed that you just had to do so much work to learn a little bit more but in this area it's like we're back in the 17th century we just study something and it's interesting and almost no extra work so it's really an exciting time to be alive working in this area um I wrote a book I wasn't intending to write a book initially on this topic but many people asked me to write a book on this topic and I ended up writing a book for archeologists and linguists and interested people the public because it really was impossible to figure out what was going on in this field from the journalistic articles and from the what had been written about it and so I thought I would try to write a book that tried to give a flavor of this field not just a flavor but describe how it's done that was not written in a jargony way and could be accessed by people who are interested in it and wanted to come to terms with it because it is very disruptive it is very challenging it is very interesting and I think it's very important thank you [Applause]
