Fly iDNA suggests strict reliance of the causative agent of sylvatic anthrax on rainforest ecosystems

Abstract Metabarcoding of invertebrate‐derived DNA (iDNA) is increasingly used to describe vertebrate diversity in terrestrial ecosystems. Fly iDNA has also shown potential as a tool for detecting pathogens. Combining these approaches makes fly iDNA a promising tool for understanding the ecology and distribution of novel pathogens or emerging infectious diseases. Here, we use fly iDNA to explore the geographic distribution of Bacillus cereus biovar anthracis (Bcbva) along a gradient from the forest within Taï National Park, Côte d'Ivoire, out to surrounding villages. We tested fly pools (N = 100 pools of 5 flies) collected in the forest (N = 25 pools), along the forest edge (N = 50 pools), and near surrounding villages (N = 25 pools) for Bcbva. Using the same iDNA, we sought to reconstruct fly and mammal communities with metabarcoding, with the aim of investigating potential links with Bcbva detection. We detected Bcbva in 5/100 fly pools and positivity varied significantly across the habitat types (forest = 4/25, edge = 1/50, village = 0/25). It was possible to culture Bcbva from all positive fly pools, confirming their positivity, while sequencing of their whole genomes revealed a considerable portion of known genomic diversity for this pathogen. iDNA generated data about the mammal and fly communities in these habitats, revealing the highest mammal diversity in the forest and considerable changes in fly community composition along the gradient. Bcbva host range estimates from fly iDNA were largely identical to the results of long‐term carcass monitoring efforts in the region. We show that fly iDNA can generate data on the geographic distribution and host range of a pathogen at kilometer scales, as well as reveal the pathogen's phylogenetic diversity. Our results highlight the power of fly iDNA for mammal biomonitoring and pathogen surveillance.