In any case, our results illustrate the usefulness of HLA typing to complement studies of mtDNA and other chromosomal markers in anthropological investigations. Akt inhibitors in clinical trials Almost all classical HLA loci are under the influence of some form of natural selection in addition to the stochastic forces – random genetic drift, demographic evolution and migration – associated with human peopling history. This has been shown through different approaches: (i) selective neutrality tests, which often reveal deviations from neutral expectations toward an excess of heterozygotes,46,48,49,51,88 although homozygous excess has also been observed; (ii) comparisons of

synonymous versus non-synonymous substitution rates, indicating an excess of amino acid replacements in the PBR of the HLA molecules;54,89 (iii) deep coalescent times of most HLA lineages, explainable by balancing selection;90

and (iv) computer simulation studies,55 more recently improved by ABC approaches to infer selection coefficients in specific situations.91 These results may be explained by our knowledge of the immune function of both class I and class II HLA molecules, the main hypothesis being that allelic diversity would have been favoured to better protect individuals in pathogen-rich environments, among other theories.92 Indirect support for this hypothesis XL184 order has been provided by Prugnolle et al.93 who found a significant correlation between HLA class I heterozygosity levels in populations and pathogen richness at the global level. However, this correlation tends to drop when Amerindian populations are not taken

into account (J.-F. Lemaître, M. Currat and A. Sanchez-Mazas, in preparation). As mentioned above, Amerindians may behave as isolated populations in which significant founder effects restrict the level of polymorphism. These populations 17-DMAG (Alvespimycin) HCl show high levels of lineage differentiation that may have been selected to cope with environmental factors. Therefore, a better investigation of the relationship between the molecular diversity of HLA alleles and the function of HLA molecules should be undertaken to confirm the hypothesis of pathogen-driven selection. On the other hand, most studies aimed at estimating selective coefficients (s) at the HLA loci showed that amino acid sites at the PBR region of HLA molecules are under weak selective constraint, as s values do not exceed a few per cent, (e.g. refs 54, 91) whereas other selected polymorphisms may reach much higher values (e.g. 10–20% for G6PD/A- relative to malaria94). Also, because it may depend on the pathogenic environment, the intensity of selection operating on the HLA loci may not be uniform across different geographic regions and may even be absent in specific geographic areas, as shown for Southwest Europe compared with Northwest Africa for HLA-DRB1.