Task 3.1.1. Theoretical aspects and simulations

We developed a spatial assignment test for inferring long-distance migration. The method performs an assignment-test to determine the populations of origin of all sampled individuals. In order to allow for non-sampled populations (so-called “ghost-populations”), the test is performed on allele frequencies that are interpolated over a geographical area using the technique of Kriging. Based on these tests it is possible to identify migrants and estimate their dispersal distance.

This information in turn is used to infer the dispersal kernel of the species. The performance of the method under different demographic and sampling scenarios has been extensively tested using simulated datasets. Based on this simulation study we identified the best sampling strategy to be used in Task T3.2.1. An R-package that performs the assignment-based test has been developed. A manuscript describing the developed method is being revised for resubmission to Heredity.

Publications:

Csilléry K., MGB Blum, OE Gaggiotti, O François, 2010, Approximate Bayesian Computation (ABC) in practice, Trends in Ecology & Evolution, 25, 410-418.