Genetic diversity is theorized to decrease in populations closer to a species’ range edge, where habitat may be suboptimal. However, generalist species capable of long-range dispersal may maintain sufficient gene flow to counteract this, though the presence of significant barriers to dispersal (e.g., large water bodies, human-dominated landscapes) may still lead to, and exacerbate, the edge effect. We used microsatellite data for 2,426 gray wolves (Canis lupus) from 24 sub-populations (groups) to model how allelic richness and expected heterozygosity varied with two measures of range edge (mainland-island position, latitude, and distance from range center) across >7.3 million km² of northern North America. We found that allelic richness and expected heterozygosity of island groups was measurably less than that of mainland groups and that these differences increased with the island’s distance to the species’ range center in the study area. Our results demonstrate how multiple axes of geographic isolation (distance from range center and island habitation) can act synergistically to erode the genetic diversity of wide-ranging terrestrial vertebrate populations despite the counteracting influence of long-range dispersal ability. These findings emphasize how geographic isolation is a potential threat to the genetic diversity and viability of terrestrial vertebrate populations even among species capable of long-range dispersal. 

This dataset is comprised of four separate sub-datasets, sourced from Carmichael et al. (2007), Musiani et al. (2007), McNay (2006), and a manuscript in progress (referred to as "MacNulty" in the data files, to be first published in Frevol et al. 2023). The raw data are comprised of sample IDs, latitude and longitude points indicating where the sample was collected or recorded, and microsatellite genetic information. To aid in re-use, the raw genetic data has also been formatted and presented for use with common population genetics software (Genepop, MICROCHECKER, Genetix, FSTAT). The dataset also includes spatial data files of the sub-populations described in the study, allelic richness and expected heterozygosity data derived from the raw and spatial data, and the R script used to create the models. 

• Carmichael LE, Krizan J, Nagy JA, et al (2007) Historical and ecological determinants of genetic structure in arctic canids. Molecular Ecology 16:3466–3483. https://doi.org/10.1111/j.1365-294X.2007.03381.x
• McNay ME (2006) Preliminary results of parentage analysis using microsatellite markers from an exploited wolf population in central Alaska. Alaska Department of Fish and Game, Division of Wildlife Conservation, Juneau, Alaska
• Musiani M, Leonard JA, Cluff HD, et al (2007) Differentiation of tundra/taiga and boreal coniferous forest wolves: genetics, coat colour and association with migratory caribou. Molecular Ecology 16:4149–4170. https://doi.org/10.1111/j.1365-294X.2007.03458.x

As this is a series of four datasets, description of methods used for collection/generation of each raw dataset are most easily found in their respective source publication.

Processing:
Removal of samples with low confidence or not enough genetic material designated by the above authors prior to creating raw data file. Data are separated into sub-populations (groups) based on prior knowledge of wolf genetics and ecology. Each dataset was run through Microchecker to identify scoring errors and the presence of null alleles, then run through Genepop to identify markers out of Hardy-Weinburg equilibrium or in linkage disequilibrium – these were removed unless they were otherwise known to not be physically linked. Allelic Richness and Expected Heterozygosity for each group were calculated in FSTAT and Genetix respectively. These values were then used for the creation of genetic diversity models. A more detailed explanation of the data processing stage can be found in the nested README files for each step and Frevol et al. 2023.

Programs and software used: QGIS, Genepop, MICROCHECKER, Genetix, FSTAT, R