April 04, 2003
Classifying Humans
Classifying humans
Francesc Calafell
Nature Genetics, volume 33 no. 4 pp 435 - 436
"Recent papers have shown the feasibility of classifying humans into categorical populations from their genotypes. How can this be reconciled with the claim that human races are biologically meaningless, and what are the implications for medical genetics projects?"
"Recent papers in Science (1) and the American Journal of Human Genetics (2) have shown that genetic polymorphisms can be used to predict the population of origin of an individual. In both reports a large number of polymorphisms were genotyped in population samples from around the world, and a model-based clustering method (3) was used by the authors to ascertain how many distinct populations can be found in the global sample and estimate the probability that an individual belongs to one of these populations."
"The algorithm used by the authors strips each sample of its original, self-reported population label and finds the groups of individuals that seem to cluster. Although the designers of the clustering method stated that these groups may not always have a clear biological interpretation (3), in both studies individuals in each group tended to fall into categories that corresponded to principal geographical (continental) divisions. At first glance, stating that the continental ancestry of each human can be identified seems to rehabilitate the discredited notion that humans can be classified typologically. Understanding what these studies really mean, however, requires a closer examination."
[Actually, nothing the author says in the remainder of the article actually undermines the notion that humans can indeed be classified typologically.]
(1) Rosenberg, N.A. et al. Science 298, 2381-2385 (2002).
(2) Bamshad, M.J. et al. Am. J. Hum. Genet. 72, 578-589 (2003).
(3) Pritchard, J.K., Stephens, M. & Donnelly, P. Genetics 155, 945-959 (2000).