A Bayesian framework to model transmissible cancer dynamics within Mya arenaria populations

CarissaMayo

University of Washington

"A Bayesian framework to model transmissible cancer dynamics within Mya arenaria populations"

Bivalve Transmissible Neoplasia (BTN) is an increasingly prevalent cancer spreading among bivalve species worldwide. BTN dynamics introduce complexities not common in many other infectious diseases due to its marine environment. Therefore, little is known about its transmission dynamics and population effects. This study develops a Bayesian framework to model BTN spread within Mya arenaria populations to address key gaps in our understanding, with a focus on statistical methodology for parameter inference and model development. We use a Bayesian compartmental modeling approach to infer and refine model parameters and leverage controlled laboratory and survey data. Laboratory data provide information on cancer cell emission, disease progression and environmental factor effects on the disease. Survey data from the field includes samples from East Coast sites that are used for fitting the model to disease progression over time in its natural environment. To capture the intricacies of BTN, our model framework builds on the traditional Susceptible-Exposed-Infectious (SEI) epidemiological model by incorporating cancer particle survivability components and the environmental effects of temperature. The Bayesian approach in our model development, implemented in STAN, provides parameter inferences and quantifies uncertainty in the results amidst limited or noisy ecological data. Future work will validate the model by comparing its predictions with 2025 survey data and conducting sensitivity analyses to identify key parameters. This statistical framework not only advances our understanding of BTN, but also demonstrates the applicability of Bayesian modeling in developing complex ecological and epidemiological compartmental models.
Additional authors:

---