Modern humans and more ancient hominins interbred many times throughout Eurasia and Africa, and the genetic flow went both ways.


Jordana Cepelewicz

Quanta Magazine
Modern humans and more ancient hominins interbred many times throughout Eurasia and Africa, and the genetic flow went both ways.

We carry genes from our ancestors’ encounters with ancient people like the Neanderthals, but the Neanderthals already carried some modern human genes from even earlier encounters with vanished groups., Olena Shmahalo/Quanta Magazine


Humans today are mosaics, our genomes rich tapestries of interwoven ancestries. With every fossil discovered, with every DNA analysis performed, the story gets more complex: We, the sole survivors of the genus Homo, harbor genetic fragments from other closely related but long-extinct lineages. Modern humans are the products of a sprawling history of shifts and dispersals, separations and reunions — a history characterized by far more diversity, movement and mixture than seemed imaginable a mere decade ago.

But it’s one thing to say that Neanderthals interbred with the ancestors of modern Europeans, or that the recently discovered Denisovans interbred with some older mystery group, or that they all interbred with each other. It’s another to provide concrete details about when and where those couplings occurred. “We’ve got this picture where these events are happening all over the place,” said Aylwyn Scally, an evolutionary geneticist at the University of Cambridge. “But it’s very hard for us to pin down any particular single event and say, yeah, we’re really confident that that one happened — unless we have ancient DNA.”

You think you’re just looking at a Neanderthal, but you’re actually looking at a mixture of Neanderthal and modern human. 

Adam Siepel, Cold Spring Harbor Laboratory

The events that do get pinned down therefore tend to be relatively recent, starting with the migration of modern humans out of Africa 60,000 years ago, during which they interacted with hominin relatives (like the Neanderthals and Denisovans) they met along the way. Evidence of interbreeding during any migrations before then, or during events that transpired earlier within Africa, has been elusive.

Now that’s starting to change. In part because of greater computational power, “we’re starting to see the next wave of methods development,” said Joshua Akey, a professor of genomics at the Lewis-Sigler Institute for Integrative Genomics at Princeton University. “And that’s allowing us to start making new inferences from the data … that the previous generation of methods couldn’t make.”

As scientists peer further back in time and uncover evolutionary relationships in unprecedented detail, their findings are complicating the narrative of human history and rescuing some formerly missing chapters from obscurity. Clues are emerging about the unexpected influence of gene flow from ancient hominins on modern human populations before the latter left Africa. Some researchers are even identifying the genetic contributions modern humans might have made to those other lineages, in a complete reversal of the usual scientific focus. Confusing and intertwined as these many effects can be, all of them shaped humanity as we now know it.

Old Humans, New Tricks

When researchers first recovered DNA from Neanderthal bones, the available techniques for making sense of it were powerful but relatively simple. Scientists compared ancient and modern sequences, tallied up shared sites and mutations, and conducted bulk statistical analyses. That’s how they discovered in 2010 that Neanderthal DNA makes up approximately 2% of the genome of people today of non-African descent, a result of interbreeding that occurred throughout Eurasia beginning 50,000-60,000 years ago. That’s also how they discovered that Denisovan DNA makes up approximately 3% of the genome of people in Papua New Guinea and Australia.

“But that kind of very simple approach isn’t very good at sorting out the complexity” of how those lost populations interacted, said John Hawks, a paleoanthropologist at the University of Wisconsin, Madison. Nor does it allow researchers to test specific hypotheses about how that interbreeding unfolded.

Population geneticists could backtrack through the DNA data to identify common ancestors from hundreds of thousands of years ago, and they could detect recent incidents of gene flow from the past few tens of thousands of years. But discerning interbreeding that occurred between those periods, from events “old enough not to be recent but young enough not to be ancient,” Hawks said, “that actually takes an extra trick.” That’s because the more recent events smear their footprints over the older ones; the DNA sequences left behind from those older events are so fragmented and mutated that they are difficult to recognize, and even more difficult to label with a date and location.

The quantitative biologist Adam Siepel and his team at Cold Spring Harbor Laboratory campus searched through contemporary and fossil DNA for signs of gene flow from modern humans into Neanderthals. Constance Brukin/CSHL

Adam Siepel, a quantitative biologist at Cold Spring Harbor Laboratory in New York, and his colleagues decided to focus on such gaps in the narrative. They were particularly interested in looking for signs of gene flow from modern humans into Neanderthals. That flow of genetic information is harder to study than the reverse, not only because of how long ago it happened, but also because there are fewer genomes to refer to: Think of all the present-day genomes at researchers’ disposal, versus the handful of Neanderthal genomes left intact, or the single genome recovered from Denisovan remains. The challenge again prompted the need for new methods.

Using one such new technique, first in 2016 and then again in a preprint posted earlier this summer, Siepel and his team found that around 3% of Neanderthal DNA — and possibly as much as 6% — came from modern humans who mated with the Neanderthals more than 200,000 years ago. The same group who gave rise to modern humans throughout the world also furnished Neanderthals with (at least a little) more DNA than the Neanderthals would later give them. “You think you’re just looking at a Neanderthal,” Siepel said, “but you’re actually looking at a mixture of Neanderthal and modern human.”

“That’s cool,” Hawks said. Such a high level of genetic admixture, he added, “is like saying 6% of the cars on the road that you see are red, but somehow you never noticed any red cars. You ought to notice that.” And yet the methods in general use had not. To Hawks, the omission suggests that there may be a lot more shared genetic material still to find even if it can’t yet be quantified accurately. More advanced techniques may change that.

More Than a One-Off

The finding also adds to the already compelling body of evidence that there were multiple migrations of modern humans out of Africa, stretching back over hundreds of thousands of years. Modern humans were thought to have evolved in Africa after the departure of Neanderthals and Denisovans, and to have remained on the continent until their well-known out-of-Africa diaspora 60,000 years ago. But recently, fossil evidence has indicated otherwise: A human jawbone in Israel, reported last year to date back to 180,000 years ago, and a skull fragment in Greece that’s even older, indicate earlier human migrations.

In fact, with that piece of skull, archaeologists may have stumbled across a possible member of the long-ago exodus that Siepel and his team inferred in their genomic study. The fossil, which was classified as Neanderthal when it was unearthed in Greece in the 1970s, was analyzed last month by the paleoanthropologist Katerina Harvati of the University of Tübingen and her colleagues. Structurally, it looked somewhat like a modern human skull, but it was estimated to be about 210,000 years old — supposedly too old to be modern at that location. (Because the structural similarities to modern skulls show up in reconstructive models of the Greek fossil, the conclusion is controversial and will probably continue to be until DNA can be recovered for a genetic study to confirm it.)

The Apdima 1 skull fossil found in Greece has many modern structural features but is 210,000 years old — too ancient to be from any of the modern humans who left Africa only 60,000 years ago. It may have come from a hypothesized earlier exodus that left no survivors. Photograph by Nicholas Thompson, ©️ Katerina Harvati, University of Tübingen

Now the DNA evidence seems to back up this revised migration narrative as well. In retrospect, “it seems quite natural,” Scally said, “to say that human populations and evolution were just as messy 200,000 years ago, and just as subdivided and structured … as they are today.”

“It makes it hard to argue that there was ever some … special evolutionary event or genetic event that triggered the evolution of humans as we know them,” he added. Humans have been continuously evolving through the mixing of varied populations for hundreds of thousands of years. (In fact, Scally posits that our species did not originally evolve from a single population in Africa, but rather from many interconnected populations spread out across the continent.)

“This is telling us, ‘Oh, this is not a weird one-off,’” Hawks said. “It’s a continuing interaction.”

Clues are emerging about the unexpected influence of gene flow from ancient hominins on modern human populations before the latter left Africa.

What is curious is that the only migration that seems to have left modern human descendants in Europe and Asia was the one from 60,000 years ago. The groups that migrated earlier apparently all died out or got absorbed into Neanderthal or other ancient populations. “If there were earlier events,” Scally said, “they left essentially no ancestry or negligible ancestry in us today.”

This could mean, he said, that “this Neanderthal legacy could be the only descendants that those people had.” Furthermore, when the Neanderthals then interbred with modern humans during later migrations, perhaps some of that DNA got mixed back into the modern human genome, embedding older signals of Homo sapiens history into the genetic material of individuals alive today.

According to Siepel’s analysis, that sort of nested mixing seems to have been exactly what happened with the Denisovans. When the team looked at the Denisovan genome, they found fragments of DNA in it from an even earlier hominin, vestiges of some population whose own genome has not been found or sequenced. It might have been Homo erectus, which split off from the ancestors of modern humans and spread across Eurasia about 1 million years ago. The contribution from this unidentified group “was at the limits of our detection power,” according to Siepel, because it constituted only about 1% of the Denisovan genome. During later interbreeding events, tiny pieces of that 1% got passed on to modern humans in Southeast Asia, Papua New Guinea and some parts of East Asia. “A small set of extremely divergent DNA sequences present in modern humans, if our analysis is correct, would have been passed through two interbreeding events,” Siepel said.

A Return to Africa

“Basically,” Akey summed it up, “the lesson is that when populations meet, they mix.” Serena Tucci, a postdoctoral researcher in Akey’s lab, said the work shows “the need that we have for more sophisticated computational approaches, for a computational framework to make inferences about our past.”

In Siepel’s case, that meant testing a vast number of hypotheses by inferring the branching inheritance patterns of various genes. Other scientists are starting to rely on different probabilistic approaches. “As computational power continues to become more sophisticated, these types of methods will become increasingly accessible and feasible to do,” Akey said. “And really, you can’t do better than these models. They use all the features of the data.”

It suggests that maybe Neanderthals actually are us. As different as they are, maybe they’re just another version of us. 

John Hawks, University of Wisconsin, Madison

Siepel now hopes to apply his approach to other elusive aspects of history. He’s particularly interested in prehistoric population dynamics on the African continent. How ancient genetic admixture events affected modern African genomes has been little studied — although a pair of researchers recently reported in PLOS Genetics that humans in Africa interbred with another ancient hominin group both before and after the ancestors of European and Asian populations split off and migrated away. By the scientists’ estimates, DNA from that unknown group now makes up somewhere between 4% and 8% of modern human ancestry.

That said, Siepel’s technique could perhaps provide deeper insights into those statistics and what they mean: For example, researchers studying how that ancient DNA made its way out of Africa into other populations might follow its trail to map out, if only sketchily, migrations as yet unknown.

“I think Africa is one of the areas that’s going to give a lot more data in the future,” said Chris Stringer, an anthropologist at the Natural History Museum in London and a member of the research team that studied the Greek fossil.

Siepel is also using his algorithm to look for signs of natural selection acting on these DNA sequences: Were ancient hominins any better or worse off for carrying more genes from modern ones? So far, his team has found no evidence for either positive or negative selection in the flow of genes from modern humans into Neanderthals 200,000 years ago, which indicates that “most of this gene flow … is just a signature of populations in contact,” according to Hawks.

“It suggests that maybe Neanderthals actually are us,” he said. “As different as they are, maybe they’re just another version of us.”

That’s something that can be studied in other species as well: Siepel has already started to look into the forces at work in the speciation of certain birds. “What we should be doing is taking these more complicated models that we have now, this messy picture … and applying that to other species,” Scally said.

Of course, inferring these population histories is a complicated process. “There is a limit to what genetics can infer, too,” Akey said. Sometimes, alternative historical scenarios have basically the same effects on the genomic record, and in those situations, even better methods of genetic analysis will be hard-pressed to squeeze answers out of the data. Still, he added, we’re a long way off from reaching that limit.

Scally agreed. “There is an enormous amount of information in human diversity today,” he said. “There’s plenty of stuff still for us to discover.”

Jordana Cepelewicz is a staff writer at Quanta Magazine who covers biology. Her writing about mathematics, neuroscience and other subjects has also appeared in Nautilus and Scientific American. Before entering the world of science reporting, Jordana did editorial work at Harper’s MagazinePolitico and Tea Leaf Nation. She graduated from Yale University in 2015 with bachelor’s degrees in mathematics and comparative literature.



Interpret the world and change it


Privacy Policy

To unsubscribe, click here.