December 22, 2010
Denisovans Bred With Humans Outside Of Africa

Razib Khan has been dropping hints that some big story about human evolution was about to break. Finally the official announcements are here and it is quite a story. "Archaic" humans separate from Neanderthals bred with some human populations and some humans alive today carry some of their genes. Is that cool or what?

Researchers have discovered evidence of a distinct group of "archaic" humans existing outside of Africa more than 30,000 years ago at a time when Neanderthals are thought to have dominated Europe and Asia. But genetic testing shows that members of this new group were not Neanderthals, and they interbred with the ancestors of some modern humans who are alive today.

Well, if two such groups are possible is there a third group waiting to be identified? In theory we should be able to detect the presence of other lost groups that inbred with humans by sequencing the genomes of every human population. Look for sequences that seem out of place. Super cheap DNA sequencing will make that possible. What secrets lurk in the genes of Andaman Islanders, the Ainu of Hokkaido Japan, the Eskimos, or the Australian Aborigines?

Fossils of these Denisovans were found in a cave in Siberia.

The journal Nature reported the finding this week. The National Science Foundation's Behavioral and Cognitive Sciences Division partially funded the research.

An international team of scientists led by Svante Pääbo at the Max Planck Institute of Evolutionary Anthropology in Leipzig, Germany, used a combination of genetic data and dental analysis to identify a previously unknown population of early humans, whom the researchers call "Denisovans." The name was taken from Denisova Cave in southern Siberia where archaeologists from the Russian Academy of Sciences recovered a bone in 2008.

The finger bone of a girl provided the needed DNA sample.

Genetic sequencing of DNA extracted from a finger bone of a 5-10-year-old girl from the cave revealed that she was neither Neanderthal nor a modern human, but shared an ancient origin with Neanderthals. The genetic analysis also showed she had a very different history since splitting from Neanderthals, the researchers concluded.

A tooth from the cave is unlike human teeth. But what about the total shape of the Denisovans? The obvious thing to try: Clone them in a human egg. If a bunch of them are brought back into existence will they start communicating with each other telepathically and try to take over the planet?

4 to 6 percent of the genes of the people of Papua New Guinea come from Denisovans.

Another type of analysis reported by the study's authors showed Denisovans contributed 4-6 percent of their genetic material to the genomes of present-day New Guineans. "They are ancestors of people in Papua New Guinea but not of the great majority of people in Eurasia," said David Reich, a geneticist at Harvard Medical School in Boston, who led the research's population genetics analysis.

Check out pictures of the tooth.

Update: Carl Zimmer's coverage in the New York Times includes someone way more knowledgeable speculating the same point as I guessed at above: there could be more interbreeding cases waiting to be found.

Dr. Bustamante also thinks that other cases of interbreeding are yet to be discovered. “There’s a lot of possibility out there,” he said. “But the only way to get at them is to sequence more of these ancient genomes.”

Is ancient genome sequencing the only way to discover evidence of interbreeding? These genetic sequences from distant relatives of humans don't stand out for other reasons? I would think whole chromosome sequencing might identify chromosomes that couldn't possibly have come from humans who left Africa in the last 100,000 years. No?

Update II: Some comments from Greg Cochran in a Gene Expression thread point out that even before the paper reported on above there were signs of homo erectus admixture in Melanesians.

There has for a long time been a suspicion that Australoids had erectus admixture.
I’ve also seen funny genetic anomalies that are probably due to this.

There were further hints this year. Long, looking at microsatellites, found evidence for one admixture that showed up in all Eurasians and another that showed up only in Melanesians. Moreover, Linda Vigilant (from Max Planck) found Long’s work interesting and said that it fit certain patterns they had seen in Melanesians. Later, in the fall, I noticed the clues in Table S48. I thought that the Denisova sample might be from the same population (from Occam’s razor), but was somewhat discouraged from this when Paabo said the Denisova pinkie was Neanderthal, as recently as two weeks ago.

As for ancient population substructure in Africa – the idea that it explained the evidence of Neanderthal admixture was silly. The idea that it might explain Denisovan admixture in New Guinea is the turducken of silly.

I would expect some genetic differences in a human population can be too complex to be the product of that population's evolution by itself. If more such genetic signatures of admixture exist they will be found. The cost of genetic sequencing is getting too cheap for these patterns to remain undetected.

On Discover Magazine's GNXP blog Greg Cochran says the Denisovans are probably homo erectus.

“Unless of course you are suggesting that Denisovans=Asian Erectines??”

Of course I am. The dates in this paper are functions of the assumed mutation rate. We have two different estimates for that, one much-used standard rate based on essentially nothing, and a recent, much lower one one based on parent-offspring rates and known mutation rates for Mendelian diseases. In the paper, they used the standard rate. Switch to the lower rate and you get population split times that fit the fossil record better in both Europe and Asia.

Yet it’s a great paper for all that.

Also see a post by John Hawks: The Denisova genome FAQ

Share |      Randall Parker, 2010 December 22 05:53 PM  Evolutionary History

razib said at December 22, 2010 6:21 PM:

i should ask nick patterson, one of the authors. he had access to andaman sequences from the india paper, so i wonder if he did a quick & dirty and found nothing? the HGDP pakistan populations should have 1/3 as much as the andman islanders because of ancient south indian, and they don't seem too. so i think it's only melanesian. oz aborigines should have it, new guinea separated from oz only ~10,000 years ago. ainu would be interesting.

Randall Parker said at December 22, 2010 7:05 PM:


I read the Ainu page in Wikipedia (linked above) and they mentioned genetic commonality with Andaman Islanders. So I wonder what groups share rare genetic sequences with the the Melanesians. Do these sequences show up in Kiri Te Kanawa or other Maori?

Radford Neal said at December 22, 2010 7:53 PM:

Relevant to the previous post too...

People don't inherit whole chromosomes from their parents. During generation of sperm or eggs, the two copies of each chromosome in a parent undergo a "crossing over" operation, so that the resulting chromosome in the sperm or egg is a combination of the two chromosomes in the parent, not one or the other of these two chromosomes. This greatly increases the variety of offspring possible, and allows good variants of genes to be separated from bad variants of other genes on the same chromosome. The benefits of sex would be much reduced if this didn't happen.

Randall Parker said at December 22, 2010 9:16 PM:


My understanding is that cross-overs are rare events. I would be happy to be corrected on this. I would be curious to know what the frequency is for cross-overs.

Radford Neal said at December 22, 2010 9:58 PM:

My understanding is that crossing over (in humans) typically happens in one or a few places on each chromosome every generation (varying, of course, with how big the chromosome is). This is supported by computations using the numbers at, taking into account that there are about 3 billion base pairs in the human genome, and 23 chromosomes. (Hence about 100 million base pairs per chromosome, with about 1% chance of recombination between bases separated by one million base pairs, yielding something like one recombination on average per chromosome in each generation.)

razib said at December 22, 2010 11:45 PM:

I read the Ainu page in Wikipedia (linked above) and they mentioned genetic commonality with Andaman Islanders. So I wonder what groups share rare genetic sequences with the the Melanesians. Do these sequences show up in Kiri Te Kanawa or other Maori?

re: ainu, the japanese has a whole have a weird cluster of D NRY haplogroups. that's what's shared with ainu. tibetans have it too. the andaman islanders might have it...but the indian populations in the HGDP population are in the normal cluster from what i can see. like the bougainville islanders they should deviate some. the polynesians should have it, but diluted. they're all 20-25 % melanesian.

Roderick Reilly said at December 23, 2010 2:40 PM:

My ex is convinced that I am one of these archaic humans and that my family is a result of intensive inbreeding.

BioBob said at December 23, 2010 2:43 PM:

As far as I am concerned, demographics = evolution. All these geneticists speculating on evolutionary dates from WAG rates of DNA change pale by comparison to the evidence from simple geometric population increase. Modern day humans 1st evolved in south China and India because that is where the largest human populations are today, and nobody can gainsay that obvious and simple evidence.

I am just waiting for the genetic evidence to come around and follow; it appears we won't have to wait all that much longer.

no said at December 23, 2010 4:10 PM:

That sounds very simplistic Bob.
What with super volcanos and meteor strikes nearly finishing off early man in the past, who's to say that where most of them reside today is where they orginated.
150 years ago, Europeans used to make up a large proportion of the worlds population due to rapid increases during the industrial revolution. That didn't make Europe the birth place of the modern form of human.

The only supprise I have with this new evidence is that the powers that be actually haven't covered it up. The idea that all humans are closely related Africans always was more political than anything else.

BioBob said at December 23, 2010 7:03 PM:

@ no said at December 23, 2010 4:10 PM: "sounds very simplistic"

yes, simple is always good - if a simple explanation is available, it often is the correct one as well. I am unaware of any supervolcano's and meteor strikes within the last 80,000 years other than those like Tambora or Krakatoa, but none of those caused all that massive a global type human population dieback. And sure, if the evolutionary history of a species is old and subject to millions of years of passing time, demographics gets pretty complicated. But no such problems arise with human evolution since we are a brand-spanking-newcomer. Certainly there was a global dieback after the 1st global spread of humans since we had incipient race formation, but that process followed the birth of humanity by some unknown period. By all means, supply some convincing explanation why the geometric population increase should be ignored. I have seen some very unconvincing ones.

Actually, there are more Europeans alive today than there ever were in the past - all 360 - 400 million of them compared to 1.3+ billion Chinese and 1.2+ billion Indians. Euro populations were never all that high compared to asian populations - look it up. ALL of the globes largest population centers are in southern asia. ALL human populations have pretty simple geometric increases up to the mid-20th century, as far as I can see, even accounting for catastrophic disease outbreaks like the black death in Europe, which generate a short term blip in that geometric increase but nothing more.

I am more than willing to be patient and wait for the science to catch up with the obvious.

Lee Reynolds said at December 24, 2010 12:24 AM:

Wow...almost a dozen comments on a story about human evolution and not a single creationist rant among them.

Lets see if I can't channel some of the crazy for a moment......

(clears throat)

"Hark! Repent! It was Lucifer who put those bones in the dirt to be found by us. He who commands all the mean and low things of the ground. He seeks to challenge our faith and lead us astray. Human beings were not bred from monkeys! We were fashioned from clay on day 6. The end times are near!"

Ok... enough crazy.

Ed Unneland said at December 24, 2010 8:42 AM:

Could one suggest that there are multiple waves from Africa? Erectus first comes from Africa and populates Asia, and evolves in situ in parallel with Homo in Africa (but differently), some becoming Denisova. An Erectus population from Asia radiates into Europe and over time becomes Neanderthalensis. Archaic Sapiens migrates from Africa to Asia and ethnically cleanses the existing populations of Erectines (with some interbreeding with local females), pushing Erectus and interbred Erectus/Sapiens into Papua, Australia and Oceana. (With some Archaic Sapiens later making the left turn after the Caucasus into Europe, where they do much the same to Neanderthalensis, but perhaps this is a later development and is done with more interbreeding and less conflict.) I'm only guessing, so completely open to correction.

John Gorter said at December 24, 2010 9:19 AM:

Given the probable several pre European invasions of Australia, which aboriginal group would you expect to see these characteristics in? On the principal that the Pitjanjatara (sic) in central Australia are there only because they were "pushed" there by subsequent invasions, would you expect these characteristics to be more prevelent in that grouping? And what about the Tasmaians, separated about 10,000 years ago?

John Gorter

Ed Unneland said at December 24, 2010 9:35 AM:

@John: given that I am rather innocent of any knowledge of the Aborigines of Australia, I should simply concede your points and I should probably refrain from kibbitzing ... I guess you are suggesting that the timescales are probably insufficient for any significant evolutionary divergence ...

Michael Gilson said at December 29, 2010 7:50 PM:

I read some fascinating information based on louse DNA in "Before the Dawn". Body lice can only survive under clothing, and the calculated time of their divergence from head lice predates Homo Sapiens. If accurate that means it was Erectus that invented clothing, even if nothing more than cloaks and or kilts. Furthermore there are two genetic groups of body lice and the time of their divergence corresponds to when some Erectus left Africa. The two groups indicate that the Sapiens who evolved in Africa later must have encountered Erectus outside of Africa, just as is claimed for these Denisovans

Phillep Harding said at December 30, 2010 11:21 AM:

Note to BioBob: If the greatest concentration of house cats were to be found in Sun City, Arizona, would you argue that housecats evolved there?

I think a better criteria is the number of diseases native animals share with humans. Not perfect, just better. Africa and the Fertile Crescent would both be contenders, I suspect.

Post a comment
Name (not anon or anonymous):
Email Address:
Remember info?

Go Read More Posts On FuturePundit
Site Traffic Info
The contents of this site are copyright ©