Roughly 60,000 years ago, the researchers argued, modern humans must have expanded from Africa and interbred with Neanderthals. The hybrid descendants passed their genes to later generations, who spread around the globe.
That hypothesis has held up well over the past decade, as paleoanthropologists have extracted more complete Neanderthal genomes from other fossils.
But Joshua Akey, a geneticist at Princeton University who carried out some of these studies, grew dissatisfied with the methods used to look for Neanderthal DNA in living people. The standard method was built on the assumption that most Africans had no Neanderthal DNA at all.
Dr. Akey and his colleagues figured out a new method, which they call IBDMix, that takes advantage of the fact that relatives share stretches of matching DNA.
Siblings, for example, share many long, identical stretches of DNA. But their children will have fewer identical segments, which will also be shorter. Distantly related cousins will have tinier matching segments that require sophisticated methods to uncover.
Dr. Akey and his colleagues figured out how to search the DNA of living humans and remains of Neanderthals for these minuscule matching segments. Then they pinpointed the segments that came from a relatively recent ancestor — and therefore were a sign of interbreeding.
The scientists searched 2,504 genomes of living humans for segments that matched those in a Neanderthal genome. When the scientists tallied up the results, the results took Dr. Akey by surprise.