Thursday, August 25, 2011

More thoughts on Paleo Hanky Panky

In an article in the American Journal of Physical Anthropology, Ghirotto and Colleagues report the results of a study of mitochondrial DNA from both early modern human fossils, Neanderthals, and living Europeans. These authors argue that their study supports a model with no Neanderthal contribution to modern human mitochondrial DNA.  This clearly contradicts the results of nuclear DNA studies, which provide strong evidence for interbreeding between Neanderthals and moderns.  What gives? Either Neanderthals and moderns got it on or they didn' one of these studies must be wrong, right?  In the future I hope to have a guest post by a geneticist about why looking at different genes can give you different answers to the same question  In the meantime, I offer my thoughts on models of modern human origins. 

First, let's not frame this as a discussion between a full replacement model versus a multiregional model.  The thing about models is that, in the word of George E.P. Box....all models are wrong. That is something that even scientists tend to forget! The  whole point of such models is to frame the logical extreme positions.  I think the image below, from a 2004 article by Osborn Pearson in Evolutionary Anthropology, illustrates this nicely. Thanks to Brett Nachman for pointing out this article.

Figure 2 from Pearon, 2004, Evolutionary Anthropology 13:145-159

The two "competing" models are on the extreme ends, and the evidence tends to be somewhere between the two. However, in this case "between the two" is shifted heavily towards the right side of the figure. The only change I made to this figure is that I drew my own box (in red) to illustrate what the current DNA evidence seems to indicate.  I would say the interbreeding was much less than 25% as Pearson's reading of the evidence allowed in 2004.  Equally important, however, is that my red box excludes 0%, because we now know with near-certainty that Neanderthal genes did make it into modern human populations.  In other words, we now have proof that a Strict Out of Africa model is wrong.  Of course, this does bring to mind the rest of the George E.P. Box quote above. 

All models are wrong. 
The practical question is: how wrong do they have to be to not be useful?