Habitat shape and population arrangement: essential aspects in landscape genetics

In current landscape genetic studies the landscape between a pair of populations is correlated with gene flow in order to identify landscape elements that facilitate or inhibit gene flow. However, in these pairwise comparisons the arrangement of populations surrounding the respective population pair is not taken into account, although this arrangement may have an influence on pairwise gene flow estimates. Because the majority of landscape genetic studies makes use of historic gene flow measures (e.g. F_ST), measured gene flow is not only the result of direct migration between populations, but it may result from indirect gene flow via surrounding populations. The arrangement of populations in a study area is strongly influenced by the shape of the habitat, which is an integral part of the landscape configuration. Here, we first demonstrate how habitat shape can influence gene flow rates, by studying the emergence of isolation-by-distance (IBD) patterns in various neutral landscapes. For this aim, we developed a population genetic model that simulates distance-weighted migration between populations that were randomly placed in habitat. We show that the emergence of certain IBD patterns strongly depends on the shape of the habitat. In a second study, we show how landscape genetic results can be improved if population arrangement is taken into account. This study was performed on an empirical genetic dataset of the wetland grasshopper Stethophyma grossum, from which we sampled all populations in a study area. In comparison to a landscape genetic analysis performed on all possible population pairs (i.e. traditional method), we show that results strongly improved when the analysis was restricted to only those population pairs that were within migration distance and also further restricted to the population pairs without any intermediate populations. From the above results, we conclude that habitat shape and population arrangement should be more generally considered in landscape genetic analysis. However, in order to achieve this goal, more emphasis has to be placed on complete sampling of all populations in a study area.