My latest Mind and Matter column discusses the debate about how
non-Africans got their 1-4% Neanderthal DNA:
So did we or didn't we? Last week saw the publication of two new
papers with diametrically opposed conclusions about whether
non-African people have Neanderthal-human hybrids among their
ancestors-a result of at least some interspecies dalliance in the
That non-Africans share 1% to 4% of their genomes with
Neanderthals is not in doubt, thanks to the pioneering work of
paleo-geneticists led by the Max Planck Institute's Svante Paabo.
At issue is how to interpret that fact. Dr. Paabo originally
recognized that there are two possible explanations, hybridization
(which got all the press) or "population substructure."
The second explanation goes like this. Around 350,000 years ago,
when the ancestors of Neanderthals first left Africa and began to
populate Eurasia, the African population from which they came would
not have been genetically homogeneous. The early Neanderthals would
therefore have been more closely related to some of their African
cousins than to others.
That would have continued to be true for long afterward, because
the genes of northern Africans would not have mixed fully with
those of southern Africans. It might still have been true when the
ancestors of modern Eurasians left Africa, about 65,000 years ago.
In other words, northern Africans might still have had genetic
similarities to Neanderthals, compared with southern Africans, left
over from more than 350,000 years ago-and they might have brought
that genetic similarity with them to Europe.
It is this argument that Dr. Andrea Manica and Dr. Anders
Eriksson have tested with mathematical modeling and found plausible. The model uses over 100
populations in Africa to represent a diversity of geographical
races. "Hopefully," says Dr. Manica, "Everyone will become more
cautious before invoking hybridization, and start taking into
account that ancient populations differed from each other probably
as much as modern populations do."
Dr. Paabo, however, along with Dr. David Reich of Harvard
University and others, has now come to
the opposite conclusion. Measuring the length of genetic
sequences that remain intact in the DNA shared by modern people and
Neanderthals, these researchers infer that "the last gene flow from
Neanderthals into Europeans likely occurred 37,000-86,000 years
ago, and most likely 47,000-65,000 years ago." That is much too
late for the substructure theory: Neanderthals were a separate
species long before then. But because this time frame matches the
period of modern people's exodus from Africa, it strongly supports
the notion of recent interbreeding.
Moreover, as Drs. Paabo and Reich point out, there is intriguing
archeological evidence for where and when the interbreeding might
have happened. Around 100,000 years ago, modern humans occupied
caves in what is now Israel. By 70,000 years ago, in cooler times,
the Neanderthals were in the area. By 50,000 years ago, the moderns
had come to the Middle East to stay. These long epochs of
to-and-fro migration must have given opportunities for sexual
The population-substructure argument also struggles to explain
an even larger genetic overlap between yet another extinct human
form, the "Denisovans," and people from Southeast Asia and
Australasia, presumably originating from an eastern exodus out of
Africa. With the addition of this case, the substructure argument
requires two different, distinct human groups with different
samples of pre-modern-human genetic material (Neanderthal and
Denisovan) to have migrated at different times-and to have ended up
in the same parts of the world as ancient populations whose genes
they happen to share.
It is probably more plausible (and newsworthy) to imagine that
when modern people spread around the Indian Ocean, they too
encountered a distantly related human species and dallied with them
under the palms.