Distinguishing between population bottleneck and population subdivision by a Bayesian model choice procedure

Authors

  • BENJAMIN M. PETER,

    1. Computational and Molecular Population Genetics (CMPG), Institute of Ecology and Evolution, University of Bern, Baltzerstrasse 6, CH-3012 Bern, Switzerland
    2. Swiss Institute of Bioinformatics, 1015 Lausanne, Switzerland
    Search for more papers by this author
  • DANIEL WEGMANN,

    1. Computational and Molecular Population Genetics (CMPG), Institute of Ecology and Evolution, University of Bern, Baltzerstrasse 6, CH-3012 Bern, Switzerland
    Search for more papers by this author
    • Present address: Department of Ecology and Evolutionary Biology, University of California Los Angeles, Los Angeles, CA 90095, USA.

  • LAURENT EXCOFFIER

    1. Computational and Molecular Population Genetics (CMPG), Institute of Ecology and Evolution, University of Bern, Baltzerstrasse 6, CH-3012 Bern, Switzerland
    2. Swiss Institute of Bioinformatics, 1015 Lausanne, Switzerland
    Search for more papers by this author

Laurent Excoffier, Fax: +41 31 631 48 88; E-mail: laurent.excoffier@iee.unibe.ch

Abstract

Although most natural populations are genetically subdivided, they are often analysed as if they were panmictic units. In particular, signals of past demographic size changes are often inferred from genetic data by assuming that the analysed sample is drawn from a population without any internal subdivision. However, it has been shown that a bottleneck signal can result from the presence of some recent immigrants in a population. It thus appears important to contrast these two alternative scenarios in a model choice procedure to prevent wrong conclusions to be made. We use here an Approximate Bayesian Computation (ABC) approach to infer whether observed patterns of genetic diversity in a given sample are more compatible with it being drawn from a panmictic population having gone through some size change, or from one or several demes belonging to a recent finite island model. Simulations show that we can correctly identify samples drawn from a subdivided population in up to 95% of the cases for a wide range of parameters. We apply our model choice procedure to the case of the chimpanzee (Pan troglodytes) and find conclusive evidence that Western and Eastern chimpanzee samples are drawn from a spatially subdivided population.

Ancillary