Human population structure and demographic history using genetic markers

The evolutionary history of the human species has generated complex patterns of population structure and linkage disequilibrium (non-random associations of alleles at different loci or LD). The understanding of these patterns is crucial to two of the most important challenges facing biomedical science today: the identification of disease predisposing genes and prediction of variable drug reactions. The genetic variation revealed by these endeavours can also illuminate the underlying population historical processes. Here, I illustrate each of these applications: first, by assessing the demographic context of cultural change in the British Isles. Y chromosome variation indicates that the Viking age invasions left a significant paternal legacy (at least in Orkney), while the Neolithic and Iron Age cultural transitions did not. In contrast, mitochondrial DNA and X chromosome variation indicate that one or more of these pre-Anglo-Saxon revolutions had a major effect on the maternal genetic heritage of the British Isles. Second, I provide conclusive evidence that diverse demographic histories produce strikingly different patterns of association. Elevated LD extends an order of magnitude further in the Lemba, a Bantu-Semitic hybrid population, than in the putative parental populations. A significant relationship between allele-frequency differentials in the parental populations and the Lemba LD demonstrates that it is admixture-generated. Third, I demonstrate that the genetic structure inferred in a heterogeneous sample using neutral markers (a) shows ethnic labels to be inaccurate descriptions of human population structure, and (b) predicts drug metabolising profiles, defined by the distribution of drug metabolising enzyme variants. Thus the trade-off between therapeutic response and adverse drug reactions will differ between different sub-clusters. Assessment of genetic structure during drug trials is therefore, like the empirical evaluation of each population’s pattern of LD, a necessity