Simulation and Verification for Computational Modelling of Signalling Pathways

Modelling of the dynamics of biochemical reaction networks typically proceeds by solving ordinary differential equations or stochastic simulation via the Gillespie algorithm. More recently, computational methods such as process algebra techniques have been successfully applied to the analysis of signalling pathways. One advantage of these is that they enable automatic verification of the models, via model checking, against qualitative and quantitative temporal logic specifications, for example, "what is the probability that the protein eventually degrades?". Such verification is exhaustive, that is, the analysis is carried out over all paths, producing exact quantitative measures. In this paper, we give an overview of the simulation, verification and differential equation approaches to modelling biochemical reaction networks. We discuss the advantages and disadvantages of the respective methods, using as an illustration a fragment of the FGF signalling pathway