Population Bottlenecks and Patterns of Human Polymorphisms
Jody Hey and Eugene Harris
Mol Biol.Evol.16(10):1423-1426. 1999

(Saira Khan)

Hey and Harris develop a new statistical model for examining bottlenecks among human populations unlike the traditional bottleneck model used in population genetics (Nei, Maruyama, and Chakraborty 1975). Human population genetics are more concerned with population size expansion by analyzing mtDNA. They cite Fay and Wu (1999) for recent publication on human population genetics which shows that "nuclear genes and mtDNA are not both consistent with simple historical models.

The recent model are adequate in explaining population expansion, they do not fit with the historical facts that have been discovered considering the mtDNA and Y-chromosome data. The authors try to develop a model that best describes all of the data in which the human ancestral populations became smaller and then later became larger. There are three concerns in trying to develop such a model:

1. "Tajima’s D statistic is not ideal for fitting models that depart from a constant historical population size.

2. Genetic bottleneck models are complex, and properly include at least six parameters (population sizes for before, after, and during the bottleneck; duration of the bottleneck; time since the bottleneck; and mutation). It is usually reasonable to reduce these to five parameters by scaling both mutation and time to the population sizes, but bottleneck models will still surpass the cope for inference that is available in most summarization of many data sets.

3. The mtDNA and Y-chromosome lack recombination and are expected to be subject to effects of an interaction of linkage and selection that reduces effective population size and increases the number of low-frequency polymorphisms".

This expression can be used to provide a different view of bottleneck effect. The effect of a bottleneck on the polymorphism frequency distribution for a sample of n=10 DNA sequences was plotted. The distribution of the autosomal gene and that for a mitochondrial gene are shaped differenlty and as expected showing the population decrease (first step of a bottleneck) and then a population expansion (second step in bottleneck).

The bottleneck model, and the model developed by Hey and Harris "have many parameters, and the scope for fitting all of them with any one data set will be limited. However, the effect described offers the hope that complex models of historical population size change can be fit to miltilocus data sets.