The question of whether modern humans made love or war with our ancestors has swung back and forth over decades of often acrimonious debate. At present, most researchers trying to read prehistory in our genomes believe that we contain no trace of species past and that we are all descended from a group that left Africa within the past 100,000 years and replaced all other humans, such as Neandertals, without interbreeding.

Those who favor the alternative view feel that the issue is moving their way, however. “Things are coming to the surface that don’t fit that model” of a single recent migration out of Africa, says Murray Cox of the University of Arizona. Cox believes that he and his colleagues have found the clearest sign so far that modern humans mated with Homo erectus, a species that originated about two million years ago that many believe has a place on the lineage leading to ourselves.

The case for the all-conquering Africans is based mainly on studies of the Y chromosome and the mitochondrion, an energy-generating structure within the cell that has its own small genome and passes down the female line. All modern variation in these sequences traces back to Africa and shows relatively recent common ancestors. (The absence of obviously hybrid fossils also supports this idea.)

But, Cox says, there are plenty of other places for genes from our relatives to hide. Studying the genetics of modern human populations, he and his colleagues stumbled across a region of the X chromosome, called RRM2P4, that shows large differences between people in different places—a sign of an ancient genetic split. To pin down the gene’s origins, the researchers sequenced some 250 individuals, half from Africa, the rest from China, Central Asia, the Basque Country of southwestern Europe, and the Pacific Islands.

Translating the differences between gene sequences into a date for their divergence, the researchers conclude that the various forms of RRM2P4 last shared a common ancestor about two million years ago—around when H. erectus migrated from Africa into Asia. And the oldest variant seems to have originated in Asia, where it is now almost exclusively found.

The combination of great antiquity and Asian origins is compelling, Cox believes. Fossils suggest that H. erectus may have survived in Asia up until about 30,000 years ago, overlapping with modern humans by about 15,000 years. “This gene is most common exactly where you find Homo erectus fossils,” says Cox, who described the finding in the January edition of Genetics. The team also used new statistical techniques to show that the gene is significantly more likely to have arisen in Asia than in Africa. The possibility that
H. erectus and modern humans interbred is all the more surprising, he adds, because most researchers think there is no evidence for our having swapped genes with the more closely related Neandertals.

Other researchers note that many factors can create deep genetic divisions among human groups. No one disputes that the large majority of our genome has a recent African origin. But that dominance makes it difficult to tell what might be the legacy of a small amount of breeding with other lineages and what might be chance. Across the whole genome, some groups will show large differences even if they all left Africa at about the same time, explains geneticist Peter Underhill of Stanford University. “When you see these outliers, are they really emblematic of something dramatic, or are they just the extremes of the normal distributions? To me, it’s still an open question,” he says.

Alternatively, the Asian form of RRM2P4 might have been present in a group that left Africa but later died out in that continent. Or the researchers might simply have not looked at enough Africans to find it; in fact, the DNA sequence that Cox and his colleagues think came from H. erectus existed in one African. This finding suggests that Asian RRM2P4 probably did originate in Africa, argues Peter Forster, who studies what has become known as archaeogenetics at Anglia Ruskin University in Cambridge, England. “It spoils the story considerably to find this gene in an African,” he says, adding that “I don’t see it as convincing evidence” of ancient mixing between humans.

Cox counters that the lone African’s RRM2P4 sequence is identical to the Asian group, suggesting that this person probably descended from recent Asian immigrants. The team also has preliminary data on a second DNA region that shows an equally ancient split, where one group seems to be exclusively Asian.

No one gene can settle the matter, says geneticist Rosalind Harding of the University of Oxford. Nevertheless, several studies suggest that the human genome records some strikingly deep splits between populations. Her group, for example, has found that part of the gene for hemoglobin seems very old. If discoveries of ancient sequences continue to stack up, archaeogeneticists might eventually be convinced that we all have a little erectus in our blood.