A population geneticist looks back in time in search of human origins.

When and where anatomically modern humans evolved is arguably one of the most fundamental scientific questions. The issue also has philosophical and possibly even moral implications because it influences our definition of humanity. But I became involved in the subject for much more prosaic reasons. I was trying to make sense of the distributions among human populations of different versions of genes that imbue resistance to infectious diseases. It struck me that attempting to do this without a clear understanding of humans' past demography was bound to end in a muddle.

Despite decades of research, the origin of modern humans is still hotly debated. In a recent paper, Laurent Excoffier and his colleagues provide the first formal statistical evaluation of the likelihood for the various schemes that have been proposed (N. J. R. Fagundes et al. Proc. Natl Acad. Sci. USA 104, 17614–17619; 2007). They conclude that a recent expansion from a single African origin is better supported by the current geographical spread of human genes than a multi-regional scenario. The multi-regional hypothesis proposes that modern humans hybridized with archaic humans, such as Homo erectus, as they spread.

This result may seem unsurprising because most genetic evidence points to an African origin some 60,000 years ago with no or negligible hybridization with archaic humans. However, there is a twist. By far the best supporting evidence for hybridization between modern and archaic humans has been the observation that, looking back, the amount of time it takes to reach the most recent common ancestor of some genes largely predates the age of our species. The extensive simulations in this paper debunk that argument by demonstrating that such cases can arise if modern humans had a recent and single African origin.

Discuss this paper at http://blogs.nature.com/nature/journalclub