DNA metabarcoding reveals that coyotes in New York City consume wide variety of native prey species and human food

Coyote (Canis latrans) Macronutrient Consumption and Diet Relative to Seasonality and Urbanization

Diet selection informs the health, fitness, and behavior of wild predators. Due to assumptions that vertebrate prey contains similar compositions of macronutrients (i.e., protein, carbohydrates, and lipids), whole prey items traditionally define carnivore diets. However, increasing evidence suggests that prey differ in terms of their macronutrient compositions, particularly relative to body size. Furthermore, omnivorous predators, like coyotes (Canis latrans), integrate both prey and nonprey diet items whose macronutrient compositions vary. This is particularly important in urbanized systems, which introduce or alter the distributions of prey (e.g., domestic pets) and nonprey (e.g., ornamental plants) foods in ways that contribute to carnivore diet selection and human-wildlife coexistence. We assessed the macronutrient composition of coyote diets seasonally and relative to urbanization in the Phoenix Metropolitan Area, AZ, USA. We collected coyote scats in the field and assessed their macronutrient compositions using values gathered from the literature, as well as the volumetric composition of diet items found in coyote scats. We then assessed the macronutrient composition of coyote diets in geometric space using the geometric framework of nutrition. We observed that the macronutrient composition of coyote diets was similar between moderately and less urbanized sites, particularly in the spring-summer season. However, coyote macronutrient consumption differed seasonally, with coyotes eating more nonprotein energy relative to protein energy when carbohydrate-rich mesquite (Prosopis spp.) was more available in the fall-winter. Our results suggest that the seasonal availability and macronutrient composition of foods contribute to coyote diets. Macronutrients directly translate to energy and subsequent animal physiology and behavior. Our findings therefore advance our understanding of coyote behavior, particularly in ways that support human-wildlife management in anthropogenic areas.

Morphological and Genetic Assessments of Coyote Diet in Qualla Boundary, North Carolina, Show Interaction with Humans

Throughout the 20th century, coyotes (Canis latrans) expanded from their historical geographic range west of the Mississippi River to a current range of almost all of North America. Over the course of this expansion, coyotes have demonstrated diverse and variable omnivorous diets that change with the food resources available. This study examined the stomach contents of 25 coyotes in an area where they are relatively new, the Qualla Boundary in North Carolina, to better understand the diets of coyotes in this area. A combination of morphological identification and DNA barcoding was used to characterize the stomach contents of coyotes. Both plant and animal material were identified from anthropogenic and natural sources, the latter including native mammals. This study provides one example of the breadth and flexibility of coyote diets and helps build an understanding of how coyotes can adapt to new conditions.

Exploring the potential effects of forest urbanization on the interplay between small mammal communities and their gut microbiota

Urbanization significantly impacts wild populations, favoring urban dweller species over those that are unable to adapt to rapid changes. These differential adaptative abilities could be mediated by the microbiome, which may modulate the host phenotype rapidly through a high degree of flexibility. Conversely, under anthropic perturbations, the microbiota of some species could be disrupted, resulting in dysbiosis and negative impacts on host fitness. The links between the impact of urbanization on host communities and their gut microbiota (GM) have only been scarcely explored. In this study, we tested the hypothesis that the bacterial composition of the GM could play a role in host adaptation to urban environments. We described the GM of several species of small terrestrial mammals sampled in forested areas along a gradient of urbanization, using a 16S metabarcoding approach. We tested whether urbanization led to changes in small mammal communities and in their GM, considering the presence and abundance of bacterial taxa and their putative functions. This enabled to decipher the processes underlying these changes. We found potential impacts of urbanization on small mammal communities and their GM. The urban dweller species had a lower bacterial taxonomic diversity but a higher functional diversity and a different composition compared to urban adapter species. Their GM assembly was mostly governed by stochastic effects, potentially indicating dysbiosis. Selection processes and an overabundance of functions were detected that could be associated with adaptation to urban environments despite dysbiosis. In urban adapter species, the GM functional diversity and composition remained relatively stable along the urbanization gradient. This observation can be explained by functional redundancy, where certain taxa express the same function. This could favor the adaptation of urban adapter species in various environments, including urban settings. We can therefore assume that there are feedbacks between the gut microbiota and host species within communities, enabling rapid adaptation.

Assessing springtime vertebrate prey of sympatric mesopredators in the southeastern United States using metabarcoding analysis

Coyotes (Canis latrans) colonized the eastern United States over the last century and formed a 3-species predator guild with bobcats (Lynx rufus) and gray foxes (Urocyon cinereoargenteus) across much of the southeastern United States. Diets among the three species vary along with respective impacts on game species such as white-tailed deer (Odocoileus virginianus) and wild turkeys (Meleagris gallopavo). To determine predation impacts on vertebrate prey and dietary overlap in consumption of prey items, we assessed diets of coyote, bobcat, and gray fox during spring, coinciding with white-tailed deer fawning and wild turkey nesting and brood rearing. We sampled across three sites along the Savannah River in South Carolina from mid-May through mid-June of 2020-2021. We collected 180 scat samples along 295.9 kilometers (71.1-122.4 km/site) of unpaved secondary roads and used DNA metabarcoding to determine vertebrate diet items. We identified predator species of scat using DNA metabarcoding and species-specific mtDNA fragment analysis (153 were coyote, 20 bobcat, and seven gray fox). Overall, we found evidence that two species, coyote and bobcat, consumed deer while all three consumed turkeys. Frequency of deer in the diet varied across sites for coyotes from 62-86% and wild turkey was present with a frequency of occurrence of 9% for coyotes, 5% for bobcats, and 14% for gray fox. Vertebrate diet specialization was evident across predator species with high frequency of deer in coyote diets, rabbits and small mammals in bobcat diets, and herpetofauna in gray fox diets. During deer fawning and wild turkey nesting and brood rearing, dietary overlap appears to be mediated by disparate selection of prey items, which reduced competition among coyotes, bobcats, and gray foxes. Use of DNA metabarcoding may augment our understanding of dietary preferences within this predator guild by providing increased resolution of diet composition among important game species.

An experimental design for obtaining DNA of a target species and its diet from a single non-invasive genetic protocol

Next-generation sequencing technology has enabled accurate insights into the diet of wildlife species. The protocols for faecal sample collection and DNA extraction for diet analysis have differed from those focusing on target species, even in most studies combining questions on both aspects. We designed an experiment to evaluate two protocols using 11 parameters and select a single one that will generate both target species (Asiatic wild ass, Equus hemionus, in Israel) and diet DNA, as an effective strategy to minimise time, effort, and cost without hampering efficiency. In Protocol A, we swabbed the outer surface of faecal boluses and extracted DNA using a Stool Kit, while for Protocol B, we homogenised faecal matter from inside the bolus followed by extraction using a Powersoil Kit. Protocol A performed significantly better for four parameters, which included, for the target species, microsatellite amplification success and the quantity of the GAPDH gene; and for its diet, the number of exact sequence variants (ESVs) obtained at genus level and plant genus richness. However, there was no significant difference in the amplification success of sex-linked and plant markers, total reads at genus level, number of genera obtained and plant genus composition. Although we chose Protocol A, both protocols yielded results for the target species and its diet, demonstrating that one single protocol can be used for both purposes, although a pilot study is recommended to optimise the protocol for specific systems. This strategy may also be useful for studies combining target species and their gut microbiome and parasitic load.