Hybrid stochastic simulation of rule-based polymerization models

Abstract

Modeling and simulation of polymer formation is an important field of research not only in the material sciences but also in the life sciences due to the prominent role of processes such as actin filament formation and multivalent ligand-receptor interactions. While the advantages of a rule-based description of polymerizations has been successfully demonstrated, no efficient simulation of these mostly stiff processes is currently available, in particular for large system sizes. We present a hybrid stochastic simulation approach, in which the average changes of highly abundant species due to fast reactions are deterministically simulated while for the remaining species with small counts a rule-based simulation is performed. We propose a nesting of rejection steps to arrive at an approach that is efficient and accurate. We test our method on two case studies of polymerization.