Identification of the spatial pattern of biodiversity is important for understanding underlying biological, physiological or demographical processes. Delineation of genetic boundaries (i.e. detection of areas of sharp change in allele frequencies) is potential approach to describe such patterns. In the recent time, there is a growing interest in applications of spatial analyses to identify genetic discontinuities and an increasing number of statistical methods implemented in freely available software. In this project, we evaluate some of the existing methods by its performances for both empirical and simulated datasets. We included some of the Bayesian spatial clustering methods (TESS, BAPS and Geneland) and two other methods (Wombling and Monmonier’ s algorithm). Preliminary results show that Bayesian clustering methods perform better than two other methods, both on simulated and different types of empirical data.

Biology,Agriculture