Wilson, Paul J

Fishers (Pekania pennanti) were extirpated from many parts of Ontario in the early 20th century, but as of the early 2000s the species had recolonized most of its historical range. While the primary population genetic structure of fishers in central and eastern Ontario has not changed drastically over the past ten years, we did find evidence of increased secondary structure and a reduction in northward movement from southeastern Ontario, a site of recent immigration from the Adirondacks in northern New York. This may be indicative of a reduction in density and thus in density-dependent migration, or it may be a consequence of the population reaching equilibrium following a period of rapid expansion associated with recolonization. We also observed no variation within central and eastern Ontario at 14 of 15 candidate functional loci we screened, suggesting possible directional or stabilizing selection and a lack of adaptive potential.

Author Keywords: fisher, functional genes, Ontario, Pekania pennanti, population genetics, recolonization

Habitat loss and fragmentation can disrupt population connectivity, resulting in small, isolated populations and low genetic variability. Understanding connectivity patterns in space and time is critical in conservation and management planning, especially for wide-ranging species in northern latitudes where habitats are becoming increasingly fragmented. Wolverines (Gulo gulo) share similar life history traits observed in large-sized carnivores, and their low resiliency to disturbances limits wolverine persistence in modified or fragmented landscapes - making them a good indicator species for habitat connectivity. In this thesis, I used neutral microsatellite and mitochondrial DNA markers to investigate genetic connectivity patterns of wolverines for different temporal and spatial scales. Population genetic analyses of individuals from North America suggested wolverines west of James Bay in Canada are structured into two contemporary genetic clusters: an extant cluster at the eastern periphery of Manitoba and Ontario, and a northwestern core cluster. Haplotypic composition, however, suggested longstanding differences between the extant eastern periphery and northwestern core clusters. Phylogeographic analyses across the wolverine's Holarctic distribution supported a postglacial expansion from a glacial refugium near Beringia. Although Approximate Bayesian computations suggested a west-to-east stepping-stone divergence pattern across North America, a mismatch distribution indicated a historic bottleneck event approximately 400 generations ago likely influenced present-day patterns of haplotype distribution. I also used an individual-based genetic distance measure to identify landscape features potentially influencing pairwise genetic distances of wolverines in Manitoba and Ontario. Road density and mean spring snow cover were positively associated with genetic distances. Road density was associated with female genetic distance, while spring snow cover variance was associated with male genetic distance. My findings suggest that northward expanding anthropogenic disturbances have the potential to affect genetic connectivity. Overall, my findings suggest that (1) peripheral populations can harbour genetic variants not observed in core populations - increasing species genetic diversity; (2) historic bottlenecks can alter the genetic signature of glacial refugia, resulting in a disjunct distribution of unique genetic variants among contemporary populations; (3) increased temporal resolution of the individual-based genetic distance measure can help identify landscape features associated with genetic connectivity within a population, which may disrupt landscape connectivity.

Author Keywords: conservation genetics, Holarctic species, landscape genetics, peripheral population, phylogeography, wolverine

Variation in habitat quality and disturbance levels can strongly influence a species' distribution, leading to spatial variation in population density and influencing population dynamics. It is therefore critical to understand how density can lead to variability in demographic responses for effective conservation and recovery of species. My dissertation illustrates how density and spatial familial networks can be integrated together to gain a better understanding of the influence of density on population dynamics of boreal caribou. First, I created an analytical framework to assess results from empirical studies to inform spatially-explicit capture-recapture sampling design, using both simulated and empirical data from noninvasive genetic sampling of several boreal caribou populations in Alberta, Canada, which varied in range size and estimated population density. Analysis of the empirical data indicated that reduced sampling intensity had a greater impact on density estimates in smaller ranges, and the best sampling designs did not differ with estimated population density but differed between large and small population ranges. Secondly, I used parent-offspring relationships to construct familial networks of boreal caribou in Saskatchewan, Canada to inform recovery efforts. Using network measures, I assessed the contribution of individual caribou to the population with several centrality measures and then determined which measures were best suited to inform on the population demographic structure. I found substantial differences in the centrality of individuals in different local areas, highlighting the importance of analyzing familial networks at different spatial scales. The network revealed that boreal caribou in Saskatchewan form a complex, interconnected familial network. These results identified individuals presenting different fitness levels, short- and long-distance dispersing ability across the range, and can be used in support of population monitoring and recovery efforts. Finally, I used a spatial capture-recapture analytical framework with covariates to estimate spatial density of boreal woodland caribou across the Saskatchewan Boreal Plains, and then reconstructed parent-offspring relationships to create a familial network of caribou and determined whether spatial density influenced sex-specific network centrality, dispersal distance, individual reproductive success, and the pregnancy status of females. I show that caribou densitygreatly varied across the landscape and was primarily affected by landscape composition and fragmentation, and density had sex-specific influences on dispersal distance, reproductive success, and network centrality. The high density areas reflected good-quality caribou habitat, and the decreased dispersal rates and female reproductive output suggest that these remnant patches of habitat may be influencing demographic responses of caribou.

Author Keywords: boreal caribou, density, familial networks, population dynamics, rangifer tarandus caribou, spatial capture-recapture

The structure and composition of the landscape can facilitate or impede gene flow, which can have important consequences because genetically isolated groups of individuals may be prone to inbreeding depression and possible extinction. My dissertation examines how landscape structure influences spatial patterns of genetic differentiation and diversity of American marten (Martes americana) and Canada lynx (Lynx canadensis) in Ontario, Canada, and provides methodological advances useful for landscape geneticists. First, I identified the effects of map boundaries on estimates of landscape resistance, and proposed a solution to the bias: a buffer around the map boundary. Second, I assessed the sensitivity of a network-based estimate of genetic distance, conditional genetic distance, to incomplete sampling. I then used these landscape genetic tools in a pairwise, distance-based analysis of 653 martens genotyped at 12 microsatellite loci. I evaluated whether forest management in Ontario has influenced the genetic structure of martens. Although forest management practices had some impact, isolation by distance best described marten gene flow. Our results suggest that managed forests in Ontario are well connected for marten and do not impede marten gene flow. Finally, I used a site-based analysis of 702 lynx genotyped at 14 microsatellite loci to investigate spatial patterns of genetic diversity and differentiation at the trailing (contracting) edge of the lynx distribution in Ontario. I analyzed harvest records and found that the southern edge of lynx range has contracted by >175 km since the 1970s. I also found that neutral genetic diversity decreased towards the trailing edge, whereas genetic differentiation increased. Furthermore, I found strong correlations between gradients of lynx genetic structure and gradients of climate and land cover in Ontario. My findings suggest that increases in winter air temperature, decreases in snow depth, and loss of suitable habitat will result in further loss of genetic diversity in peripheral populations of lynx. Consequently, the adaptive potential of lynx populations on the southern range periphery could decline. In conclusion, my dissertation demonstrates the varying influences that contemporary landscape structure and climate gradients can have on genetic diversity and differentiation of different species.

Author Keywords: Circuitscape, genetic network, landscape genetics, Lynx canadensis, Martes americana, range shift

Species ranges are changing and the rate at which the climate is warming is faster than anything previously seen in the past, consequently species will need to adapt quickly, track the climate or perish. Cold adapted terrestrial species are the most vulnerable, because they are limited by the availability of land at the cold edge of their range. This means that many alpine, boreal and polar species essentially have nowhere to go as the climate warms. Habitat generalists are widely distributed across the globe and are highly adaptable to anthropogenic change. Our future biodiversity may only consist of several habitat generalists. The Canada lynx (Lynx canadensis) is a boreal species that has limited range expansion potential at the cold end of its range and its range has already contracted by 40%. The lynx has nowhere to go as climate warming progresses in this current century. Therefore, understanding the causes of its range contraction could enlighten us on conservation and management strategies that we might undertake as climate warms. My analyses indicated that the Canada lynx seems to have tracked the habitat that it is adapted to in more northern homogenous boreal forests and the bobcat (Lynx rufus), a habitat generalist, has simply replaced it in the south.

Author Keywords: Anthropogenic Change, Competition, Connectivity, Lynx canadensis, Lynx rufus, Range change

Mountain ranges are often thought to restrict movement of wildlife, yet previous studies evaluating the role of the Rocky Mountains as a dispersal barrier for Canada lynx (Lynx canadensis) have been contradictory. Our study uses neutral microsatellite loci to evaluate the role of the Rocky Mountains as a barrier to gene flow for lynx. Although lynx exhibited low genetic differentiation, we detected a limited effect of the mountains. Furthermore, we inferred the role played by landscape variables in gene flow (genetic differentiation predicted by landscape resistance). Limited gene flow most strongly related to resistance from physical factors (low snow cover and elevation), rather than other topographic and ecological factors (high terrain roughness, low forest cover, low habitat suitability, and geographic distance). Structural connectivity was a relatively poor predictor of functional connectivity. Overall, the Rockies represent an area of reasonably high functional connectivity for lynx, with limited resistance to gene flow.

Author Keywords: Canada lynx, connectivity, gene flow, genetic structure, landscape genetics, Rocky mountains

The release of domestic organisms to the wild is considered a threat to

biodiversity because the introduction of domestic genes through interbreeding can

negatively impact wild conspecifics via outbreeding and local extinction. In North

America, captive American mink (Neovison vison) are frequently escaping into the wild,

yet the impact of these events on the functional genetic diversity of wild mink populations

is unclear. I characterized domestic and wild mink in Ontario at 17 trinucleotide

microsatellites located in functional genes thought to be associated with traits affected by

domestication. I found low functional genetic diversity, as only 4 of 17 genes were

variable and of those four there was little evidence of allele frequency differences

between domestic and wild mink. Using redundancy analysis and a spatial analysis of

principal components on the four variable loci (AR, ATN1, IGF-1, and TOB1) I found no

evidence to suggest domestic release events are affecting functional genetic diversity of

free-ranging mink at the set of markers assessed.

Author Keywords: American mink, domestication, functional gene, introgression, Neovison vison

Introgressive hybridization between northern (Glaucomys sabrinus) and southern flying squirrels (G. volans) has been observed in some areas of Canada and the USA. However, existing molecular markers lack the resolution to discriminate late-generation introgressants and describe the extent to which hybridization influences the Glaucomys gene pool. I report the first North American flying squirrel (genus Glaucomys) functionally annotated de novo transcriptome assembly with a set of 146,621 high-quality, annotated putative species-diagnostic SNP markers. RNA-sequences were obtained from two northern flying squirrels and two southern flying squirrels sampled from Ontario, Canada. I reconstructed 702,228 Glaucomys transcripts using 193,323,120 sequence read-pairs, and captured sequence homologies, protein domains, and gene function classifications. These genomic resources can be used to increase the resolution of molecular techniques used to examine the dynamics of the Glaucomys hybrid zone.

Author Keywords: annotation, de novo transcriptome, flying squirrels, high-throughput sequencing, hybridization, single nucleotide polymorphisms

Polar bears are generally described as solitary, but features of their life cycles and

habitats regularly necessitate interaction. Effective conspecific assessment, including accurate

recognition and discrimination, likely confers benefits, especially to females accompanied by

dependent young. Individuals in the Southern (SH) and Western (WH) Hudson Bay

subpopulations are ideal for studying polar bear social behaviours because of the prolonged

high densities of the ice-free season. First, I looked outside family groups to model their fine

scale sociospatial organization on land. Capture locations were more likely to correspond to

family groups when there were fewer neighbouring bears, when a greater proportion of

neighbours were female, and when the focal individual and neighbours were significantly

related. Second, I looked within the family group to assess offspring recognition. Of 288

offspring in 207 family groups captured in the SH subpopulation from 1999 through 2013, only

one case of adoption (of a singleton) was observed.

Author Keywords: Adoption, Kin Recognition, Logistic Regression, Maternity Analysis, Social Discrimination, Sociospatial

To better understand species' resilience to climate change and implement solutions, we must conserve environments that maintain standing adaptive genetic variation and the potential generation of new beneficial alleles. Coding trinucleotide repeats (cTNRs) providing high-pace adaptive capabilities via high rates of mutation are ideal targets for mitigating the decline of species at risk by characterizing adaptively significant populations. Ultimately, adaptive genetic information will inform the protection of biological diversity below the species level (i.e., "Evolutionarily Significant Units" or "ESUs"). This dissertation investigates cTNRs within candidate genes to determine their prevalence and influence under selection in North American mammals. First, I evaluated the potential for somatic mosaicism in Canada lynx (Lynx canadensis), and found that tissue-specific mosaicism does not confound cTNR genotyping success in lynx. Second, I assessed a selection of clock gene cTNRs across characterized mammals and found that these repeats are abundant and highly variable in length and purity. I also identified preliminary signatures of selection in 3 clock gene cTNRs in 3 pairs of congeneric North American mammal species, highlighting the importance of cTNRs for understanding the evolution and adaptation of wild populations. I further evaluated the influence of selection on the NR1D1 cTNR within Canada lynx sampled across Canada using environmental correlation, where I estimated the variation in NR1D1 cTNR alleles explained by environmental and spatial variables after removing the effects of neutral population structure. Although most variation was explained by neutral structure, environment and spatial patterns in eastern lynx populations significantly explained some of the variation in NR1D1 alleles. To examine the role of island populations in the generation and distribution of adaptive genetic variation, I used 14 neutral microsatellites and a dinucleotide repeat within a gene linked to mammalian body size, IGF-1, and found that both genetic drift and natural selection influence the observed genetic diversity of insular lynx. Finally, I estimated the divergence dates of peripheral lynx populations and made recommendations towards the conservation of Canada lynx; high levels of genetic differentiation coupled with post-glacial colonization histories and patterns of divergence at cTNR loci suggest at least 4 ESUs for Canada lynx across their range.

Author Keywords: adaptation, Canada lynx, candidate genes, coding trinucleotide repeats, evolution, natural selection