Patterns of nucleotide variability within and around G6PD, a locus under positive natural selection in humans

Abstract

Some mutations at the gene coding for glucose-6-phosphate dehydrogenase (G6PD) cause a clinical condition of G6PD deficiency, however these alleles have also been shown to confer resistance to malaria based on geographical allele distributions that coincide with malaria, epidemiology, and in vitro studies. A detailed study of nucleotide variability at G6PD can shed light on the signature of selection on the human genome and can also provide insight into the natural history of human malaria. In Appendix A, patterns of nucleotide variability are described in a worldwide panel of humans at G6PD and at a neighboring locus ( L1CAM). Patterns at G6PD do not significantly differ from patterns found at other loci. Nonetheless, significant long-range linkage disequilibrium (LD) is associated with a selected G6PD deficiency allele from Africa (G6PD A-), implying that the allele is young, and that malaria is a recent agent of selection in humans (within the past 10,000 years). In Appendix B, patterns of nucleotide variability are examined in a panel from sub-Saharan Africa at G6PD and at nine loci located around G6PD. LD and reduced nucleotide variability are observed spanning a region of >1 Mb around G6PD, confirming a young age for G6PD A- and implying that the allele has been under strong selection (0.05 < s < 0.20). In Appendix C, a comparable survey of nucleotide variability was conducted in a Eurasian panel with respect to G6PD med, an independently arisen selected G6PD deficiency allele. Patterns of LD provide strong evidence for independent origins of G6PD med in two geographic regions, and that resistance to malaria arose at approximately the same time in Africa and Eurasia. In Appendix D, a panel of Kurdish Jews was studied at G6PD, at neighboring loci, and at 2 unlinked loci to determine if the remarkably high frequency of G6PDmed in this population is attributable primarily to selection or to a severe bottleneck. Patterns of nucleotide variability that are consistent with selection are observed around G6PD, while nucleotide variability at unlinked loci is typical of other non-African populations, suggesting that natural selection is largely responsible for the high frequency of G6PDmed among Kurdish Jews.