Abstract
Many high-consequence human and animal pathogens persist in wildlife reservoirs. An understanding of the dynamics of these pathogens in their reservoir hosts is crucial to inform the risk of spill-over events, yet our understanding of these dynamics is frequently insufficient. Viral persistence in a wild bat population was investigated by combining empirical data and in-silico analyses to test hypotheses on mechanisms for viral persistence. A fatal zoonotic virus, European Bat lyssavirus type 2 (EBLV-2), in Daubenton's bats (Myotis daubentonii) was used as a model system. A total of 1839 M. daubentonii were sampled for evidence of virus exposure and excretion during a prospective nine year serial cross-sectional survey. Multivariable statistical models demonstrated age-related differences in seroprevalence, with significant variation in seropositivity over time and among roosts. An Approximate Bayesian Computation approach was used to model the infection dynamics incorporating the known host ecology. The results demonstrate that EBLV-2 is endemic in the study population, and suggest that mixing between roosts during seasonal swarming events is necessary to maintain EBLV-2 in the population. These findings contribute to understanding how bat viruses can persist despite low prevalence of infection, and why infection is constrained to certain bat species in multispecies roosts and ecosystems.