why don't all white people look alike? huh??

[furnishesq]

I was surprised Cecil responded to this question. “look alike” is a very relative question. White people “look alike” compared to Asians and Africans. If an army of pod aliens cames down to invade earth, humans would “look alike” etc.

All I know is that if someone posted this question on the message board, they would be slammed by countless other posters. . .

[system]

I thought it perhaps interesting to cite to this discussion in the recent paper: David B. Goldstein and Lounès Chikhi “Human Migrations And Population Structure: What We Know and Why it Matters” Annu. Rev. Genomics Hum. Genet. 2002. 3:129-152 to illustrate and expand on the Neandertal side point made here.

The authors, in the context of discussing the importance of population genetics and current views to medical application touch upon the robustness of single loci analyses, and put that in the context of historical genetic research. They note the recent mtDNA sequencing is an interesting example of the limitations of single loci analyses, citing to a piece of literature I missed critiquing strong conclusions off of the same.

An interesting example comes from the recent sequencing of mtDNA fragments from Neanderthal remains (Krings et al. 1997). The fact that the sequences obtained were highly divergent from all human sequences ever observed does not in fact rule out interbreeding between Neanderthal and Homo sapiens in Europe. Nordborg (Nordborg 1998) showed by simulations that even with these apparently clear-cut results, the data only allows us to reject panmixia between Neanderthal and Homo sapiens. He showed that even if Neanderthal had interbred in a small but reasonable proportion with early Homo sapiens (such as a contribution of 25% some 70 KY ago), this could easily be invisible in present-day mtDNA due to the amount of drift that has taken place since then. It follows from this, that the absence of Neanderthal mtDNA lineages does not preclude such lineages in other parts of the genome. Extending his simulation Nordborg (Nordborg 2001) has argued that it is currently impossible to reject the MR model based on a single locus study. Similarly, Wall (Wall 2000) has argued that as many as 60 loci might be needed if we wish to reject the MR model.

Interesting critiques, however, our authors continue on to note that the critique itself is lacking in engagement with the present state of current knowledge on population structures (or better, it took a highly abstract view to point out a weak point in relying too strongly on such analyses). They state:

We should note however, that while Nordborg’s point is valid so far as it goes, the conclusions are less relevant to question of archaic genetic input than it might appear. In particular, Nordborg’s arguments are based on analyses of a single unstructured population. * In fact, in any reasonable version of the MR model, different geographic areas (e.g. East Asia and Europe) will be partially or wholly isolated. But as noted above, there no “divergent” mtDNA or Y chromosome lineages have been observed in any modern populations. Taking this geographic structure into account greatly increases the amount of information provided by single locus analyses because the absence of old lineages requires their independent loss in multiple, semi or wholly isolated populations. It becomes more difficult to argue that the archaic lineages were lost by chance if they must be lost repeatedly in semi-isolated populations. The more parsimonious explanation is that the archaic lineages simply weren’t there in the first place. * Clearly, these arguments require quantification, but the important point is that we must be careful not to over-extend the applicability of very specific models.

Bold emphasis in the original, italics added to highlight. Thus the theoretical door is not quite closed to archaic Homo inputs, but as a matter of overall view of the data, the door is closing rapidly with no data really holding out for any form of MR (as a form of theorizing archaic pre Hss form inputs into modern populations – the authors elsewhere note that both OoA and MR are usually poorly specified, which often confuses the issue).

They also have some interesting comments on the problems involved in current theorizing in re the settlement of Europe and the dispute over the degree of Neolithic versus Paleolithic population inputs, arguing that the data has not bee properly used (argument that the ‘age’ of haplogroups does not tell us when such groups arose in re Europe as I recall, w/o rereading that section).

[system]

Strange, the above only seems to have shown up when I went to redo the reply.