Stream Trees: A Statistical Method for Mapping Genetic Differences Between Populations of Freshwater Organisms to the Sections of Streams that Connect Them
Statistical approaches for studying the spatial distribution of genetic diversity that assume that organisms move through a two-dimensional landscape are not well suited to study populations of freshwater fish. We present a new statistical method for mapping genetic differences among populations of freshwater fish to the sections of streams that connect them. The method is useful for freshwater species that can only disperse through stream corridors and for other species that live in habitats for which there is one, and only one, corridor connecting each pair of populations (e.g., alpine organisms confined to ridge tops). The model is a simple extension of the least-squares method for constructing evolutionary trees. In this model, the genetic distances between populations are modeled as a sum of genetic distances mapped onto landscape features (e.g., stream sections). Analysis of simulated data shows that the method produces useful results with realistic amounts of data. The model was fit to empirical microsatellite data from four metapopulations of freshwater fish and showed an excellent fit in three out of four cases. Software to perform the necessary calculations is available from the authors at www.montana.edu/kalinowski.
Kalinowski, S.T., M.H. Meeuwig, S.R. Narum, and M.L. Taper. 2008. Stream trees: a statistical method for mapping genetic differences between populations of freshwater organisms to the sections of streams that connect them. Canadian Journal of Fisheries and Aquatic Sciences 65(12): 2752–2760. Online at https://doi.org/10.1139/F08-171.