Novel multievent mark-recapture model
Matthijs Hollanders and Dr. J. Andrew Royle recently developed a novel multievent mark-recapture model that accounts for infection state assignment errors. By estimating the false-negative and false-positive probabilities in the disease detection protocols, these errors can be propagated and accounted for while estimating the ecological process of interest. They used simulations and a case study with Fleay’s barred frog (Mixophyes fleayi) infected with the amphibian chytrid fungus Batrachochytrium dendrobatidis as a case study.
They found that infection prevalence was underestimated by \(\frac{1}{3}\) while the rates of gaining and clearing infections were overestimated by factors of 4–5 when state misclassification was not accounted for. This was largely due to the limited ability of swab samples to detect low-level infections of the chytrid fungus.
The research was published as an Open Access article in Methods in Ecology and Evolution, and includes code to simulate and analyse your own datasets in the Supporting Information and on GitHub.
