High genomic diversity and candidate genes under selection associated with range expansion in eastern coyote (Canis latrans) populations

Genetic diversity and family groups detected in a coyote population with red wolf ancestry on Galveston Island, Texas

Background

Hybridization can be a conservation concern if genomic introgression leads to the loss of an endangered species’ unique genome, or when hybrid offspring are sterile or less fit than their parental species. Yet hybridization can also be an adaptive management tool if rare populations are inbred and have reduced genetic variation, and there is the opportunity to enhance genetic variation through hybridization. The red wolf (Canis rufus) is a critically endangered wolf endemic to the eastern United States, where all extant red wolves are descended from 14 founders which has led to elevated levels of inbreeding over time. Red wolves were considered extirpated from the wild by 1980, but before they disappeared, they interbred with encroaching coyotes creating a genetically admixed population of canids along coastal Texas and Louisiana. In 2018, a genetic study identified individuals on Galveston Island, Texas with significant amounts of red wolf ancestry. We collected 203 fecal samples from Galveston for a more in-depth analysis of this population to identify the amount of red wolf ancestry present and potential mechanisms that support retention of red wolf ancestry on the landscape.

Results

We identified 24 individual coyotes from Galveston Island and 8 from mainland Texas with greater than 10% red wolf ancestry. Two of those individuals from mainland Texas had greater than 50% red wolf ancestry estimates. Additionally, this population had 5 private alleles that were absent in the North American reference canid populations used in this study, which included 107 southeastern coyotes, 19 captive red wolves, and 38 gray wolves, possibly representing lost red wolf genetic variation. We also identified several individuals on Galveston Island and the mainland of Texas that retained a unique red wolf mitochondrial haplotype present in the red wolf founding population. On Galveston Island, we identified a minimum of four family groups and found coyotes on the island to be highly related, but not genetically depauperate. We did not find clear associations between red wolf ancestry estimates and landscape features, such as open green space or developed areas.

Conclusion

Our results confirm the presence of substantial red wolf ancestry persisting on Galveston Island and adjacent mainland Texas. This population has the potential to benefit future red wolf conservation efforts through novel reproductive techniques and possibly through de-introgression strategies, with the goals of recovering extinct red wolf genetic variation and reducing inbreeding within the species.

Persistence and expansion of cryptic endangered red wolf genomic ancestry along the American Gulf coast

Admixture and introgression play a critical role in adaptation and genetic rescue that has only recently gained a deeper appreciation. Here, we explored the geographical and genomic landscape of cryptic ancestry of the endangered red wolf that persists within the genome of a ubiquitous sister taxon, the coyote, all while the red wolf has been extinct in the wild since the early 1980s. We assessed admixture across 120,621 single nucleotiode polymorphism (SNP) loci genotyped in 293 canid genomes. We found support for increased red wolf ancestry along a west-to-east gradient across the southern United States associated with historical admixture in the past 100 years. Southwestern Louisiana and southeastern Texas, the geographical zone where the last red wolves were known prior to extinction in the wild, contained the highest and oldest levels of red wolf ancestry. Further, given the paucity of inferences based on chromosome types, we compared patterns of ancestry on the X chromosome and autosomes. We additionally aimed to explore the relationship between admixture timing and recombination rate variation to investigate gene flow events. We found that X-linked regions of low recombination rates were depleted of introgression, relative to the autosomes, consistent with the large X effect and enrichment with loci involved in maintaining reproductive isolation. Recombination rate was positively correlated with red wolf ancestry across coyote genomes, consistent with theoretical predictions. The geographical and genomic extent of cryptic red wolf ancestry can provide novel genomic resources for recovery plans targeting the conservation of the endangered red wolf.

Coyotes in New York City Carry Variable Genomic Dog Ancestry and Influence Their Interactions with Humans

Coyotes are ubiquitous on the North American landscape as a result of their recent expansion across the continent. They have been documented in the heart of some of the most urbanized cities, such as Chicago, Los Angeles, and New York City. Here, we explored the genomic composition of 16 coyotes in the New York metropolitan area to investigate genomic demography and admixture for urban-dwelling canids in Queens County, New York. We identified moderate-to-high estimates of relatedness among coyotes living in Queens (r = 0.0–0.5) and adjacent neighborhoods, suggestive of a relatively small population. Although we found low background levels of domestic-dog ancestry across most coyotes in our sample (5%), we identified a male suspected to be a first-generation coyote–dog hybrid with 46% dog ancestry, as well as his two putative backcrossed offspring that carried approximately 25% dog ancestry. The male coyote–dog hybrid and one backcrossed offspring each carried two transposable element insertions that are associated with human-directed hypersociability in dogs and gray wolves. An additional, unrelated coyote with little dog ancestry also carried two of these insertions. These genetic patterns suggest that gene flow from domestic dogs may become an increasingly important consideration as coyotes continue to inhabit metropolitan regions.

The natural history and ecology of melanism in red wolf and coyote populations of the southeastern United States – evidence for Gloger’s rule

Background

Gloger’s rule postulates that animals should be darker colored in warm and humid regions where dense vegetation and dark environments are common. Although rare in Canis populations, melanism in wolves is more common in North America than other regions globally and is believed to follow Gloger’s rule. In the temperate forests of the southeastern United States, historical records of red wolf (Canis rufus) and coyote (Canis latrans) populations document a consistent presence of melanism. Today, the melanistic phenotype is extinct in red wolves while occurring in coyotes and red wolf-coyote hybrids who occupy the red wolf's historical range. To assess if Gloger’s rule could explain the occurrence and maintenance of melanistic phenotypes in Canis taxa, we investigated differences in morphology, habitat selection, and survival associated with pelage color using body measurements, GPS tracking data, and long-term capture-mark-recapture and radio-telemetry data collected on coyotes and hybrids across the southeastern United States.

Results

We found no correlation between morphometrics and pelage color for Canis taxa. However, we observed that melanistic coyotes and hybrids experienced greater annual survival than did their gray conspecifics. Furthermore, we observed that melanistic coyotes maintained larger home ranges and exhibited greater selection for areas with dense canopy cover and wetlands than did gray coyotes.

Conclusions

In the southeastern United States, pelage color influenced habitat selection by coyotes and annual survival of coyotes and hybrids providing evidence that Gloger’s rule is applicable to canids inhabiting regions with dense canopy cover and wetlands. Greater annual survival rates observed in melanistic Canis may be attributed to better concealment in areas with dense canopy cover such as coastal bottomland forests. We suggest that the larger home range sizes of melanistic coyotes may reflect the trade-off of reduced foraging efficiency in lower quality wetland habitat for improved survival. Larger home ranges and differential use of land cover by melanistic coyotes may facilitate weak assortative mating in eastern coyote populations, in which melanistic animals may have lower success of finding compatible mates in comparison to gray conspecifics. We offer that our observations provide a partial explanation for why melanism is relatively low (< 10%) but consistent within coyote populations throughout southeastern parts of their range.

Selection of both habitat and genes in specialized and endangered caribou

Genetic mechanisms determining habitat selection and specialization of individuals within species have been hypothesized, but not tested at the appropriate individual level in nature. In this work, we analyzed habitat selection for 139 GPS-collared caribou belonging to three declining ecotypes sampled throughout Northwestern Canada. We used Resource Selection Functions (RSFs) comparing resources at used and available locations. We found that the three caribou ecotypes differed in their use of habitat suggesting specialization. On expected grounds, we also found differences in habitat selection between summer and winter, but also, originally, among the individuals within an ecotype. We next obtained Single Nucleotide Polymorphisms (SNPs) for the same caribou individuals, we detected those associated to habitat selection, and then identified genes linked to these SNPs. These genes had functions related in other organisms to habitat and dietary specializations, and climatic adaptations. We therefore suggest that individual variation in habitat selection was based on genotypic variation in the SNPs of individual caribou, indicating that genetic forces underlie habitat and diet selection in the species. We also suggest that the associations between habitat and genes that we detected may lead to lack of resilience in the species, thus contributing to caribou endangerment. Our work emphasizes that similar mechanisms may exist for other specialized, endangered species. This article is protected by copyright. All rights reserved.

Genomic legacy of migration in endangered caribou

Wide-ranging animals, including migratory species, are significantly threatened by the effects of habitat fragmentation and habitat loss. In the case of terrestrial mammals, this results in nearly a quarter of species being at risk of extinction. Caribou are one such example of a wide-ranging, migratory, terrestrial, and endangered mammal. In populations of caribou, the proportion of individuals considered as "migrants" can vary dramatically. There is therefore a possibility that, under the condition that migratory behavior is genetically determined, those individuals or populations that are migratory will be further impacted by humans, and this impact could result in the permanent loss of the migratory trait in some populations. However, genetic determination of migration has not previously been studied in an endangered terrestrial mammal. We examined migratory behavior of 139 GPS-collared endangered caribou in western North America and carried out genomic scans for the same individuals. Here we determine a genetic subdivision of caribou into a Northern and a Southern genetic cluster. We also detect >50 SNPs associated with migratory behavior, which are in genes with hypothesized roles in determining migration in other organisms. Furthermore, we determine that propensity to migrate depends upon the proportion of ancestry in individual caribou, and thus on the evolutionary history of its migratory and sedentary subspecies. If, as we report, migratory behavior is influenced by genes, caribou could be further impacted by the loss of the migratory trait in some isolated populations already at low numbers. Our results indicating an ancestral genetic component also suggest that the migratory trait and their associated genetic mutations could not be easily re-established when lost in a population.

Cooperative Communication with Humans Evolved to Emerge Early in Domestic Dogs

Although we know that dogs evolved from wolves, it remains unclear how domestication affected dog cognition. One hypothesis suggests dog domestication altered social maturation by a process of selecting for an attraction to humans.^1-3 Under this account, dogs became more flexible in using inherited skills to cooperatively communicate with a new social partner that was previously feared and expressed these unusual social skills early in development.^4-6 Here, we comparedog (n = 44) and wolf (n = 37) puppies, 5-18 weeks old, on a battery of temperament and cognition tasks. We find that dog puppies are more attracted to humans, read human gestures more skillfully, and make more eye contact with humans than wolf puppies. The two species are similarly attracted to familiar objects and perform similarly on non-social measures of memory and inhibitory control. These results are consistent with the idea that domestication enhanced the cooperative-communicative abilities of dogs as selection for attraction to humans altered social maturation.

SambaR: An R package for fast, easy and reproducible population-genetic analyses of biallelic SNP data sets

SNP data sets can be used to infer a wealth of information about natural populations, including information about their structure, genetic diversity, and the presence of loci under selection. However, SNP data analysis can be a time-consuming and challenging process, not in the least because at present many different software packages are needed to execute and depict the wide variety of mainstream population-genetic analyses. Here, we present SambaR, an integrative and user-friendly R package which automates and simplifies quality control and population-genetic analyses of biallelic SNP data sets. SambaR allows users to perform mainstream population-genetic analyses and to generate a wide variety of ready to publish graphs with a minimum number of commands (less than 10). These wrapper commands call functions of existing packages (including adegenet, ape, LEA, poppr, pcadapt and StAMPP) as well as new tools uniquely implemented in SambaR. We tested SambaR on online available SNP data sets and found that SambaR can process data sets of over 100,000 SNPs and hundreds of individuals within hours, given sufficient computing power. Newly developed tools implemented in SambaR facilitate optimization of filter settings, objective interpretation of ordination analyses, enhance comparability of diversity estimates from reduced representation library SNP data sets, and generate reduced SNP panels and structure-like plots with Bayesian population assignment probabilities. SambaR facilitates rapid population genetic analyses on biallelic SNP data sets by removing three major time sinks: file handling, software learning, and data plotting. In addition, SambaR provides a convenient platform for SNP data storage and management, as well as several new utilities, including guidance in setting appropriate data filters. The SambaR source script, manual and example data set are distributed through GitHub: https://github.com/mennodejong1986/SambaR.

Pleistocene climate fluctuations drove demographic history of African golden wolves (Canis lupaster)

Pleistocene climate change impacted entire ecosystems throughout the world. In the northern hemisphere, the distribution of Arctic species expanded during glacial periods, while more temperate and mesic species contracted into climatic refugia, where isolation drove genetic divergence. Cycles of local cooling and warming in the Sahara region of northern Africa caused repeated contractions and expansions of savannah-like environments which connected mesic species isolated in refugia during interglacial times, possibly driving population expansions and contractions; divergence and geneflow in the associated fauna. Here, we use whole genome sequences of African golden wolves (Canis lupaster), a generalist mesopredator with a wide distribution in northern Africa to estimate their demographic history and past episodes of geneflow. We detect a correlation between divergence times and cycles of increased aridity-associated Pleistocene glacial cycles. A complex demographic history with responses to local climate change in different lineages was found, including a relict lineage north of the High Atlas Mountains of Morocco that has been isolated for more than 18,000 years, possibly a distinct ecotype.

Coyote (Canis latrans) in South America: potential routes of colonization

During the last century, the coyote (Canis latrans) has increased its distribution in Central America. Before the 1980s, it had not been recorded in Panama. New records show that coyotes have crossed the Panama Canal, indicating that continues to expand; therefore, there is a possibility that it will reach northern South America. Our objectives were to identify potential coyote colonization routes to South America, and the variables that favor its expansion. We hypothesized that habitat fragmentation benefits coyote expansion. We applied 7 algorithms to model the potential distribution of the coyote, using 196 presence records and 12 variables. The models with better performance were used to generate a consensus model. Using our consensus model and the areas with highest probability of presence, a potential colonization route was generated between Central America and northern South America. This route lies through southern Costa Rica, along the Pacific coast of Panama to the south, to the Andean mountains in northern Colombia. The variables that explained potential coyote distribution were human population density, altitude, and percentage of crops with positive influence, and tropical broadleaf forests with negative influence. These results indicate that human activities and deforestation are related to coyote distribution expansion. Actions can be implemented within the identified route to improve environmental management, in order to avoid the presence of the coyote in the ecosystems of northern South America.

A Genome-Wide Perspective on the Persistence of Red Wolf Ancestry in Southeastern Canids

The red wolf (Canis rufus), a legally recognized and critically endangered wolf, is known to interbreed with coyotes (Canis latrans). Declared extirpated in the wild in 1980, red wolves were reintroduced to northeastern North Carolina nearly a decade later. Interbreeding with coyotes was thought to be restricted to a narrow geographic region adjacent to the reintroduced population and largely believed to threaten red wolf recovery. However, red wolf ancestry was recently discovered in canids along the American Gulf Coast, igniting a broader survey of ancestry in southeastern canid populations. Here, we examine geographic and temporal patterns of genome-wide red wolf ancestry in 260 canids across the southeastern United States at over 164 000 SNP loci. We found that red wolf ancestry was most prevalent in canids sampled from Texas in the mid-1970s, although non-trivial amounts of red wolf ancestry persist in this region today. Further, red wolf ancestry was also observed in a subset of coyotes inhabiting North Carolina, despite management efforts to limit the occurrence of hybridization events. Lastly, we found no evidence of substantial red wolf ancestry in southeastern canids outside of these 2 admixture zones. Overall, this study provides a genome-wide survey of red wolf ancestry in canids across the southeastern United States, which may ultimately inform future red wolf restoration efforts.

Geographic patterns in morphometric and genetic variation for coyote populations with emphasis on southeastern coyotes

Prior to 1900, coyotes (Canis latrans) were restricted to the western and central regions of North America, but by the early 2000s, coyotes became ubiquitous throughout the eastern United States. Information regarding morphological and genetic structure of coyote populations in the southeastern United States is limited, and where data exist, they are rarely compared to those from other regions of North America. We assessed geographic patterns in morphology and genetics of coyotes with special consideration of coyotes in the southeastern United States. Mean body mass of coyote populations increased along a west-to-east gradient, with southeastern coyotes being intermediate to western and northeastern coyotes. Similarly, principal component analysis of body mass and linear body measurements suggested that southeastern coyotes were intermediate to western and northeastern coyotes in body size but exhibited shorter tails and ears from other populations. Genetic analyses indicated that southeastern coyotes represented a distinct genetic cluster that differentiated strongly from western and northeastern coyotes. We postulate that southeastern coyotes experienced lower immigration from western populations than did northeastern coyotes, and over time, genetically diverged from both western and northeastern populations. Coyotes colonizing eastern North America experienced different selective pressures than did stable populations in the core range, and we offer that the larger body size of eastern coyotes reflects an adaptation that improved dispersal capabilities of individuals in the expanding range.

Black bears alter movements in response to anthropogenic features with time of day and season

BACKGROUND

With the growth and expansion of human development, large mammals will increasingly encounter humans, elevating the likelihood of human-wildlife conflicts. Understanding the behavior and movement of large mammals, particularly around human development, is important for crafting effective conservation and management plans for these species.

METHODS

We used GPS collar data from American black bears (Ursus americanus) to determine how seasonal food resources and human development affected bear movement patterns and resource use across the Commonwealth of Massachusetts.

RESULTS

We found that though bears moved more and avoided human development during crepuscular and daylight hours than at night, bears preferentially moved through human dominated areas at night. This indicates bears were mitigating the risk of human development by altering their behavior to exploit these areas when human activity is low. This behavioral shift was most prominent in the spring, when natural foods are scarce, and fall, when energetic demands are high. We also observed a high degree of inter-individual variability among our sample of bears. Bears with a higher density of houses in their home ranges (~ 75 houses/km2) displayed less avoidance of human development than more rural bears. Furthermore, bear movement models had different explanatory variables, with preference or avoidance of a variable being dependent on the individual bear. To account for this individuality in our predictive surfaces, we projected the probability of movement for each season and time of day using a spatially weighted surface centered on each bear's home range.

CONCLUSIONS

We found that black bears in Massachusetts are operating in a landscape of fear and are altering their movement patterns to use developed areas when human activity is low. We also found seasonal and diel differences among individual bears in resource selection during movement. Accounting for these individual, seasonal, and diel differences when assessing movement for large mammals is especially important if predictive surfaces are to be used in identifying areas for conservation and management.