Incorporating spatial data into analysis of koala populations in South-East Queensland

The objective of the study was to analyse the phylogeographic patterns of koalas in SE Queensland, Australia and to compare the clustering suggested by Bayesian clustering methods with those that incorporate spatial data. Clustering methods have become commonplace for population genetic studies, allowing the identification of population boundaries with biological meaning. The incorporation of spatial data is less frequently employed, although methodologies that allow these additional data are particularly relevant to phylogeographical studies. Such methods make the assumption that individuals that are located nearby are more likely to be closely related. The koala is an arboreal marsupial found in eastern Australia. It has suffered severe population declines in southern Australia. Greater numbers exist in northern areas such as Queensland but populations here are declining and are affected by habitat destruction, population fragmentation and disease. In a study of genetic variability in south-east Queensland, we demonstrate patterns of historical division among koala populations shown by mitochondrial DNA differ to patterns of recent divisions inferred from STRUCTURE analysis of microsatellite data. Furthermore, we present results from two programs that incorporate spatial data, TESS and Geneland, demonstrating the effect of the distribution of samples on the number of inferred clusters and their distribution.