Novel polyomaviruses identified in fecal samples from four carnivore species.

Polyomaviruses are oncogenic viruses that are generally thought to have co-evolved with their hosts. While primate and rodent polyomaviruses are increasingly well-studied, less is known about polyomaviruses that infect other mammals. In an effort to gain insight into polyomaviruses associated with carnivores, we surveyed fecal samples collected in the USA from bobcats (Lynx rufus), pumas (Puma concolor), Canada lynxes (Lynx canadensis), and grizzly bears (Ursus arctos). Using a viral metagenomic approach, we identified six novel polyomavirus genomes. Surprisingly, four of the six genomes showed a phylogenetic relationship to polyomaviruses found in prey animals. These included a putative rabbit polyomavirus from a bobcat fecal sample and two possible deer-trophic polyomaviruses from Canada lynx feces. One polyomavirus found in a grizzly bear sample was found to be phylogenetically distant from previously identified polyomaviruses. Further analysis of the grizzly bear fecal sample showed that it contained anelloviruses that are known to infect pigs, suggesting that the bear might have preyed on a wild or domestic pig. Interestingly, a polyomavirus genome identified in a puma fecal sample was found to be closely related both to raccoon polyomavirus 1 and to Lyon-IARC polyomavirus, the latter of which was originally identified in human saliva and skin swab specimens but has since been found in samples from domestic cats (Felis catus).

Complex evolutionary history of felid anelloviruses.

Anellovirus infections are highly prevalent in mammals, however, prior to this study only a handful of anellovirus genomes had been identified in members of the Felidae family. Here we characterise anelloviruses in pumas (Puma concolor), bobcats (Lynx rufus), Canada lynx (Lynx canadensis), caracals (Caracal caracal) and domestic cats (Felis catus). The complete anellovirus genomes (n = 220) recovered from 149 individuals were diverse. ORF1 protein sequence similarity network analysis coupled with phylogenetic analysis, revealed two distinct clusters that are populated by felid-derived anellovirus sequences, a pattern mirroring that observed for the porcine anelloviruses. Of the two-felid dominant anellovirus groups, one includes sequences from bobcats, pumas, domestic cats and an ocelot, and the other includes sequences from caracals, Canada lynx, domestic cats and pumas. Coinfections of diverse anelloviruses appear to be common among the felids. Evidence of recombination, both within and between felid-specific anellovirus groups, supports a long coevolution history between host and virus.

Improving the accessibility and transferability of machine learning algorithms for identification of animals in camera trap images: MLWIC2.

Motion-activated wildlife cameras (or "camera traps") are frequently used to remotely and noninvasively observe animals. The vast number of images collected from camera trap projects has prompted some biologists to employ machine learning algorithms to automatically recognize species in these images, or at least filter-out images that do not contain animals. These approaches are often limited by model transferability, as a model trained to recognize species from one location might not work as well for the same species in different locations. Furthermore, these methods often require advanced computational skills, making them inaccessible to many biologists. We used 3 million camera trap images from 18 studies in 10 states across the United States of America to train two deep neural networks, one that recognizes 58 species, the "species model," and one that determines if an image is empty or if it contains an animal, the "empty-animal model." Our species model and empty-animal model had accuracies of 96.8% and 97.3%, respectively. Furthermore, the models performed well on some out-of-sample datasets, as the species model had 91% accuracy on species from Canada (accuracy range 36%-91% across all out-of-sample datasets) and the empty-animal model achieved an accuracy of 91%-94% on out-of-sample datasets from different continents. Our software addresses some of the limitations of using machine learning to classify images from camera traps. By including many species from several locations, our species model is potentially applicable to many camera trap studies in North America. We also found that our empty-animal model can facilitate removal of images without animals globally. We provide the trained models in an R package (MLWIC2: Machine Learning for Wildlife Image Classification in R), which contains Shiny Applications that allow scientists with minimal programming experience to use trained models and train new models in six neural network architectures with varying depths.