Speaker
Description
Pharmacology and toxicology studies often use compartmental ordinary differential equation (ODE) mathematical models to understand how a chemical compound distributes and exerts its effects throughout the body. Such models require experimental data to inform parameterisation. However, such data is often costly and not always available and thus we need to identify new ways to parameterise such ODE models under data-limited conditions - the aim of this work is to address this challenge. In this talk, I will explain a novel methodology that can be effectively used to parameterise biologically based ODE models. It involves a stepwise approach, including: generating stochastic Petri net (SPN) qualitative data; scaling the qualitative SPN data into quantitative synthetic data; and using this synthetic data to obtain model parameters and their associated distributions. I will demonstrate this methodology against case examples to illustrate its feasibility and reliability, including a model for the regulation of thyroid hormone levels in the brain. I will also describe troubleshooting methods to improve accuracy of this approach, including: varying the number of initial Petri Net tokens; deploying a scaling approach; and how to flex according to the nature of the experimental data provided. Through this exploration, I will illustrate how you can enhance the robustness and precision of this methodology when applied to more complex systems.
