A day after the paper by Peter and Slatkin, a new paper has appeared on the arXiv dealing with the problem of detecting directionaliy in human migration patterns. This seems to be purely methodological, so no new insights on human history to report.
A new approach to estimate directional genetic differentiation and asymmetric migration patterns
Lisa Sundqvist, Martin Zackrisson, David Kleinhans
In the field of population genetics measures of genetic differentiation are widely used to gather information on the structure and the amount of gene flow between populations. These indirect measures are based on a number of simplifying assumptions, for instance equal population size and symmetric migration. Structured populations with asymmetric migration patterns, frequently occur in nature and information about directional gene flow would here be of great interest. Nevertheless current measures of genetic differentiation cannot be used in such systems without violating the assumptions. To get information on asymmetric migration patterns from genetic data rather complex models using maximum likelihood or Bayesian approaches generally need to be applied. In such models a large number of parameters are estimated simultaneously and this involves complex optimization algorithms. We here introduce a new approach that intends to fill the gap between the complex approaches and the symmetric measures of genetic differentiation. Our approach makes it possible to calculate a directional component of genetic differentiation at low computational effort using any of the classical measures of genetic differentiation. The approach is based on defining a pool of migrants for any pair of populations and calculating measures for genetic differentiation between the populations and the respective pools. The directional measures of genetic differentiation can further be used to calculate asymmetric migration. The procedure is demonstrated with a simulated data set with known migration pattern. A comparison of the estimation results with the migration pattern used for simulation suggests, that our method captures relevant properties of migration patterns even at low migration frequencies and with few marker loci.