Read our COVID-19 research and news.

Human evolution

DNA from this finger bone, resting on top of an artist's reconstruction of the hand it came from, revealed the existence of the Denisovans.


Our ancestors may have mated more than once with mysterious ancient humans

It looked like an ordinary finger bone. But when researchers sequenced its DNA in 2010, they uncovered the existence of a group of ancient humans no one had seen before: the Denisovans. Then came an even bigger surprise. Some modern humans also carry Denisovan DNA, meaning that at some point in the ancient past, Denisovans and modern humans mated and had children. Now, a new study concludes that all that free love had some dark consequences, including male offspring that were likely sterile.

In the absence of much fossil evidence, the best way to study Denisovans is through the genes they left behind in modern humans. So population geneticists Sriram Sankararaman at the University of California (UC), Los Angeles, and David Reich at Harvard University sifted through 257 genomes of present-day people from 120 non-African populations around the world. (Africans, whose ancestors didn’t leave Homo sapiens’s original home, do not have any Denisovan heritage.) They confirmed an earlier finding that among humans living today, people from Papua New Guinea, Australia, and other parts of Oceania have the most Denisovan ancestry, between 3% and 6% of their genomes. This compares with about 2% from Neandertals for all non-African genomes.

Sankararaman and Reich found another hot spot of Denisovan ancestry in an unexpected place: South Asia. “It’s about 10% of what we see in the Oceanians,” Sankararaman explains. That’s quite a small contribution—which allowed it fly under the radar in previous studies—but it’s more than researchers expected to find based on their best models of population mixing. East Asians, in turn, have more Denisovan ancestry than Europeans but less than South Asians, the team reports today in Current Biology.

There are a few potential explanations for how modern humans ended up with Denisovan genes, Reich says. It’s possible that Denisovans and modern humans mated only once, presumably while H. sapiens were on their way to Australia and Papua New Guinea. After that, Denisovan-related modern humans mated with modern humans whose ancestors had never met Denisovans, and different populations ended up with different proportions of the archaic genes. “On the other hand, it’s entirely possible, and I think plausible, that what you’re seeing in India is evidence of mixture of ancient South Asian ancestors with local Denisovans,” that is, an entirely separate mating event than the one that led to Denisovan ancestry in Oceania, Reich says. In fact, he explains, “the data could be consistent with as many as three independent mixture events with Denisovans.”

Although it’s difficult to determine which model is right without more data from both fossils and modern human genomes, “I don’t think it’s that hard to imagine that there were multiple admixture events,” says Joshua Akey, a population geneticist at the University of Washington, Seattle, who was not involved in the study. He led a team that recently identified as many as five mixing events between modern humans and Neandertals.

Another mystery is exactly when H. sapiens hooked up with Denisovans. Fossil evidence from Denisova Cave suggests that the ancient species lived there from as early as 170,000 years ago to at least 50,000 years ago, and who knows when they might have occupied the rest of Asia. That gives Denisovans a lot of time to run into other kinds of humans and potentially have babies with them.

But Sankararaman and Reich knew that with each successive generation, the chains of Denisovan DNA would become shorter and shorter, as the two parents’ genomes broke apart and recombined to form their children’s genomes. So they counted backward, starting from the comparatively long chains of Denisovan DNA in Oceanic populations. They did the same for their Neandertal ancestry, finding that modern humans mixed with Denisovans 5000 to 10,000 years after they mated with Neandertals. Reich and Sakararaman had already estimated that Neandertals and modern humans mated approximately 50,000 years ago, which would put the Denisovan interbreeding event—at least the one that contributed to Oceanic populations—at about 40,000 years ago.

The fact that both mixing events occurred around the same time paints an intriguing picture of what H. sapiens got up to as they spread around the world, says Rasmus Nielsen, a population geneticist at UC Berkeley who wasn’t involved in the study. “It suggests a scenario of extensive, and almost free, matings” between modern humans and just about any other closely related species we met.

But there were consequences to the love fest. Some of them were positive, such as a Denisovan gene inherited by Himalayan populations that reduces hemoglobin levels in the blood, allowing people to thrive at extremely high altitudes. Sankararaman and Reich also found evidence that Papua New Guineans have a Denisovan gene that may help them detect very subtle scents. Still, they also noticed that no trace of Denisovan ancestry remained on modern humans’ X chromosome or on the genes that express themselves in the tissue of the testes. That’s a pattern scientists see in many interspecies hybrids, and it’s a sign that when the two species get together and make babies, their male offspring are infertile. “That’s the barrier that keeps the species from mixing,” Sankararaman explains. Male Denisovan–H.sapiens hybrids, in other words, represented a genetic dead end. The traces of Denisovan ancestry researchers see today had to have been passed down by female hybrids. So, if you have Denisovan ancestry, you can thank your ancient grandmothers.

*Update, 28 March, 5:12 p.m.: The photo caption in this article has been updated for clarity.