Each pie represents the percentage of the slow (white), intermediate (grey), and fast (black) acetylators in the extended panel, inferred from genotype data as described in Materials and Methods. Inset: map showing the frequency of slow (white) and fast (black) haplotypes as inferred based on biochemical data .
Table S1 has the population frequencies. From the paper:
From a phylogeographic perspective, the consistency of the networks across different populations means that all the major haplotypes had originated prior to the differentiation of the study populations. However, the out-of-Africa bottleneck predicts lower diversity levels in non-African compared to African populations, and this is not observed at the NAT2 locus. There is no marked tendency of any of the derived NAT2 haplotypes to be continent-specific.
Several features of the data generated by us and by others , – suggest that selection acted on multiple slow-causing variants.
In conclusion, our data favour the hypothesis that a selective pressure drove at least three slow-causing variants to the frequencies observed today, but do not allow us to discriminate between balancing selection and directional selection on multiple standing variants.
PLoS ONe doi: 10.1371/journal.pone.0003136
Multiple Advantageous Amino Acid Variants in the NAT2 Gene in Human Populations
Francesca Luca et al.
Genetic variation at NAT2 has been long recognized as the cause of differential ability to metabolize a wide variety of drugs of therapeutic use. Here, we explore the pattern of genetic variation in 12 human populations that significantly extend the geographic range and resolution of previous surveys, to test the hypothesis that different dietary regimens and lifestyles may explain inter-population differences in NAT2 variation.
The entire coding region was resequenced in 98 subjects and six polymorphic positions were genotyped in 150 additional subjects. A single previously undescribed variant was found (34T>C; 12Y>H). Several aspects of the data do not fit the expectations of a neutral model, as assessed by coalescent simulations. Tajima's D is positive in all populations, indicating an excess of intermediate alleles. The level of between-population differentiation is low, and is mainly accounted for by the proportion of fast vs. slow acetylators. However, haplotype frequencies significantly differ across groups of populations with different subsistence.
Data on the structure of haplotypes and their frequencies are compatible with a model in which slow-causing variants were present in widely dispersed populations before major shifts to pastoralism and/or agriculture. In this model, slow-causing mutations gained a selective advantage in populations shifting from hunting-gathering to pastoralism/agriculture. We suggest the diminished dietary availability of folates resulting from the nutritional shift, as the possible cause of the fitness increase associated to haplotypes carrying mutations that reduce enzymatic activity.