Most computational models in systems biology are population models that trace the copy numbers or concentrations of biomolecules participating in the cellular process under study. However, a cell is highly structured, containing several compartments that can dynamically change their size and geometry as well as mechanical structures such as actin filaments or microtubules.
In order to assess the accuracy of such simple population models we are developing a Brownian dynamics-based simulator of biochemical reactions. More specifically, we aim at introducing a new meso-scale simulation regime that is more abstract than coarse-grained molecular dynamics simulation but more accurate than standard reaction-diffusion models or simple population models. This should enable us to simulate reaction dynamics in an realistic in vivo environment at physiological time scales on the order of minutes.