Researchers Identify a Third Ancient Human Lineage That Interbred With Our Ancestors in Africa

Scientists already knew modern humans interbred with Neanderthals and Denisovans. That's settled science, backed by actual recovered genomes from those groups' bones. Now a team largely based at the University of California, Berkeley says the evidence points to a third group, one whose DNA is still detectable in living Africans today, according to Ars Technica.

There's a catch. Nobody has found a skeleton, a skull, or any physical fossil from this supposed third lineage. No genome has been recovered from any specimen. Researchers are calling it a "ghost lineage" because all they have is an inference from statistical patterns in modern human DNA, not a body.

How They Found It

The method relies on something called an ancestral recombination graph: a computational reconstruction of the genome's history, base by base, tracking when a given DNA sequence likely first appeared and how often it got shuffled around through recombination, the process where chromosome pairs swap segments.

According to Ars Technica, the Berkeley team looked for a specific signature: clusters of DNA sequences that look unusually old compared to the rest of the genome, but that also show fewer signs of recombination than you'd expect from something that old. That combination is the tell. It suggests DNA that sat outside the main human population for a long stretch, then got reintroduced later, meaning it hasn't had as much time inside modern humans to get shuffled by recombination.

That pattern, the researchers argue, points to interbreeding between early modern humans and a separate, unidentified hominin group, and it happened in Africa, before any modern humans migrated out.

Why This Isn't New, Exactly

Hints of unexplained ancient ancestry in African genomes have shown up in prior research for years, according to Ars Technica's own account. What's different now is the analytical horsepower. Newer computational techniques let researchers squeeze more detail out of existing genomic datasets without needing a new physical discovery.

This isn't a new fossil dig. It's a new way of reading data that was, in many cases, already collected. The finding is a product of better math, not a shovel in the ground.

What This Doesn't Tell Us

A fair skeptic would ask: how do you know this is a distinct hominin lineage and not just an artifact of the model, or ordinary genetic variation among modern human subpopulations that predates known migrations? That's a legitimate question, and it's the same kind of question scientists had to answer before Neanderthal and Denisovan admixture became accepted fact.

The difference with Neanderthals and Denisovans is that scientists eventually got physical remains and could sequence actual genomes from those groups, cross-checking the statistical predictions against real biological material. That confirmation step doesn't exist yet for this "ghost lineage." Everything here rests on inference from modern DNA and a model of how recombination should behave over time.

That doesn't make the finding wrong. Ancestral recombination graph methods have a track record, and the specific signature the Berkeley team describes, old-looking sequences with unusually low recombination rates, is a sensible thing to look for if you're hunting for late-introduced ancient DNA. But "sensible signature in a statistical model" and "proven ancient human relative" are two different confidence levels, and readers deserve to know which one this is.

What Happens Next

The obvious next step is the same one that turned Neanderthal DNA from a hypothesis into a fact: find a body. Without fossil remains or an actual ancient genome from this population, the "ghost lineage" stays an inference, not a discovery in the traditional sense. Whether African fossil sites eventually yield remains that match the DNA signature the Berkeley team describes is the open question that will determine if this finding gets upgraded from statistical inference to established anthropology.