Abstract
Many insects and other animals host heritable endosymbionts that alter host fitness and reproduction. The prevalence of facultative endosymbionts can fluctuate in host populations across time and geography for reasons that are poorly understood. This is particularly true for maternally transmitted Wolbachia bacteria, which infect roughly half of all insect species. For instance, the frequencies of several wMel-like Wolbachia, including wMel in host Drosophila melanogaster, fluctuate over time in certain host populations, but the specific conditions that generate temporal variation in Wolbachia prevalence are unresolved. We implemented a discrete generation model in the new R package symbiontmodeler to evaluate how finite-population stochasticity contributes to Wolbachia fluctuations over time in simulated host populations under a variety of conditions. Using empirical estimates from natural Wolbachia-Drosophila systems, we explored how stochasticity is determined by a broad range of factors, including host population size, maternal transmission rates, and Wolbachia effects on host fitness (modeled as fecundity) and reproduction (cytoplasmic incompatibility; CI). While stochasticity generally increases when host fitness benefits and CI are relaxed, we found that a decline in the maternal transmission rate had the strongest relative impact on increasing the size of fluctuations. We infer that non- or weak-CI-causing strains like wMel, which often show evidence of imperfect maternal transmission, tend to generate larger stochastic fluctuations compared to strains that cause strong CI, like wRi in D. simulans. Additional factors, such as fluctuating host fitness effects, are required to explain the largest examples of temporal variation in Wolbachia. The conditions we simulate here using symbiontmodeler serve as a jumping-off point for understanding drivers of temporal and spatial variation in the prevalence of Wolbachia, the most common endosymbionts found in nature.