Two Ghost Ancestors Found Hiding in Living DNA

Every person alive is carrying a little of someone who was never dug up. No skull, no jawbone, no scrap of frozen finger bone in a Siberian cave. Just a faint statistical shadow, threaded through the genome, left by a population that vanished so long ago it never left a body we’ve found. And now, for the first time, researchers can point to where in your DNA that shadow falls.

A team at the University of California, Berkeley and Johns Hopkins University reports that modern humans interbred with at least two archaic relatives who left no sequenced genome behind. They have mapped the leftover DNA, and put rough dates on when the mixing happened.

The trick is a method they call TRACE, short for Tracking Archaic Contributions via ARG Estimation, and the intuition behind it is oddly simple. Genealogies, Priya Moorjani and her colleagues reasoned, preserve a record of the past. Every stretch of your genome has its own family tree, an ancestral recombination graph or ARG in the jargon, reaching back through time, and where a chunk of DNA came from a very distant relative, that tree has an unusually long, deep branch. Find the branches that reach back too far, and you have found ancestry from a population that split off from us long before the rest of the genome did. No ancient bone required. The living genome is the archive.

That matters because the bones have nearly run out. Only six high-coverage archaic genomes have ever been sequenced, four Neanderthals and two Denisovans, all of them from Eurasia, and DNA older than a million years almost never survives outside permafrost.

So the team built TRACE as a statistical model, ran it first on simulated genomes where they knew the right answer, then let it loose on 503 real genomes from around the world. In the simulations it did well, recovering known introgression with high precision. On the real data it recovered the Neanderthal and Denisovan DNA everyone already knew about, which was the point: a method that finds the known signal can be trusted with the unknown one.

And there was an unknown one. Buried in the deep branches was ancestry from a lineage that matches neither Neanderthal nor Denisovan, a ghost whose genome no one has ever read. “We were actually able to find and map genomic locations in modern humans that are from this ghost lineage and show that this ghost ancestry is in all modern humans, not only in Africans,” says Yulin Zhang, a Berkeley graduate student and co-first author.

The Ghost Is in Everyone

That last part is the surprise. Earlier hints of ghost ancestry had shown up mainly in African populations, and it was an open question whether it was an African-only story. TRACE says no. The ghost DNA turns up in Africans and non-Africans alike, at roughly the same dose, about 0.5 to 1 per cent of each person’s genome. Comparable, as it happens, to how much Neanderthal you carry. The even spread is itself a clue: it means the mixing happened before the ancestors of today’s non-Africans made their final move out of Africa, some 50,000 years ago, carrying the ghost DNA with them as they went.

When did the ghost lineage split from ours? The model puts it deep, around 800,000 years back, about when the Neanderthal and Denisovan branch was peeling away too. And worth keeping in mind here: the ghost is not a fossil. It’s an inference, a population defined entirely by the trace it left, never observed directly. The researchers can say it was there and roughly when; they can’t yet say who it was, though Middle Pleistocene Homo groups in Africa are a plausible fit for the timing.

The second lost ancestor is stranger still, and it came in through a side door. Call it the super-archaic lineage: so old, roughly 1.8 million years, that it predates the common ancestor of us, Neanderthals and Denisovans. It seems to have bred with Denisovans first, and only reached us second-hand, riding along inside Denisovan DNA into the genomes of people in Oceania, who carry the most Denisovan ancestry of anyone. “The super-archaic finding is particularly exciting because it reveals genetic contributions from a human lineage that lived over a million years ago, despite the absence of any sequenced DNA from that population,” says Arjun Biddanda, a postdoctoral researcher at Johns Hopkins and the other co-first author. Only a sliver of it survives in living people, a conservative lower bound, but its signature, unusually deep and diverged, is hard to mistake for anything else. One candidate for who it was, given the age, is Homo erectus.

Ghosts in the Deserts

There are stretches of the human genome long thought to be off-limits to archaic DNA, so-called deserts, swept clean of Neanderthal and Denisovan sequence and often read as regions where being distinctly Homo sapiens mattered enough that foreign DNA was selected out. TRACE finds ghost ancestry sitting right there in those deserts. In one desert on chromosome 7, home to a much-studied gene involved in speech and language, ghost DNA reaches a frequency of over 13 per cent. That reframes the deserts: perhaps they are not Homo sapiens sanctuaries so much as places where Neanderthal and Denisovan DNA specifically did not sit well, while older ghost DNA was tolerated just fine.

Much of the archaic DNA, ghost included, is not scattered at random either. It clusters in regions tied to immunity and metabolism, the parts of the genome that handle new pathogens and new foods. “Interbreeding with other human groups introduced new genetic variation, providing additional raw material for natural selection,” says Moorjani. Beneficial bits could be kept and spread; the rest thinned out over the generations.

None of this is settled. TRACE leans on reconstructed genealogies, and with real data its recall is lower than with the simulated kind, so the proportions it reports are floors rather than final counts. The number of ghost mixing events, exactly when and where they happened, whether the super-archaic DNA reached us only through Denisovans or by some other route too: all still open. What TRACE offers is a way to keep asking without waiting on a fossil that may never turn up.

The bigger shift is to the shape of the family tree, or rather the shape it is not. For a long time the story of us has been drawn as a branching tree, lineages splitting and going their separate ways. What the genomes keep saying, in Neanderthals and Denisovans and now in two populations we have never laid eyes on, is that the branches kept touching. Reaching back across the gaps, mixing, again and again.

Frequently Asked Questions