DNA Markers and Genetic Variation in the Human Species

The first attempt at using genetic markers for an-thropological purposes goes back to 1918, when Hirsz-feld and Hirszfeld (1918-1919) showed that ABO blood groups varied with the ethnic origin of population samples. Using ABO, MN, and Rh blood group data, Boyd (1950) showed one could assign some populations to their ethnic group of origin. Only rarely is an indi-vidual locus sufficiently informative that conclusions about ethnic origins can be drawn. Nevertheless, evo-lutionary theory predicts that by accumulating infor-mation from many genes one may be able to recon-struct the evolutionary history of a group of populations. With this in mind, Cavalli-Sforza nd Ed-wards (1964) attempted to reconstruct a tree of descent of human ethnic groups using five loci for which infor-mation was then available for 15 populations. A new method specifically designed for reconstruction f phy-logenies and a classical statistical method of reduction of dimensions (the analysis of principal components) both showed that all populations clustered in a geo-graphically sensible way. A figure showing similar clus-terings was published in the 1964 Cold Spring Harbor Symposium volume which, like the present Sympo-sium, was dedicated to humans (Cavalli-Sforza et al. 1965, Fig. 2). Tree structures can depict two kinds of informa-t ion- the existing genetic relationships among the populations (both qualitative and quantitative) and the evolutionary history represented by the relationships of the populations to the origin, or root, of the tree. The two methods that were developed in the early 1960s (least-squares fitting of an additive tree, and minimum path) did not generate a root. The root in those early trees was placed about midway between the two most