Estimating the population structure of brown bears in eastern Hokkaido based on microsatellite analysis

Estimating the genetic structure of a population is important for the conservation and management of wildlife. In the present study, our aim was to estimate the genetic structure of the brown bear (Ursus arctos) population in eastern Hokkaido by performing a Bayesian clustering analysis. To accomplish this goal, we used 15 microsatellites to generate genotypic data from tissue samples collected from 646 bears between 1996 and 2008. Using this genotypic data and the geographic locations where the bears were captured, GENELAND analysis detected six subpopulations. Based on the genotypic data, the STRUCTURE analysis revealed three subpopulations. As inferred from the GENELAND analysis, the core zones of the subpopulations (G-a through G-f) were located in the Shiranuka Hills (G-a), the northern area of the Shiranuka Hills (G-b), the eastern slope of the Daisetsuzan Mountains (G-c), the northern slope of the Akan Mountain Range (G-d), the Shiretoko Peninsula (G-e), and Akkeshi District (G-f). The STRUCTURE analysis indicated that G-b and G-d were influenced by gene flow from other subpopulations. National routes, towns, and farm fields were considered to have formed the distribution boundaries among the subpopulations. A high level of genetic differentiation was not observed among the six subpopulations, with the exception of G-f (Fst = 1.35–0.176, Ds = 0.246–0.349), which showed a geographically discontinuous distribution. We suggest that the loss of forest areas through future regional development and road building should be avoided to facilitate gene flow in brown bears in Hokkaido.