Title:A scalable method for finding irreducible state-spaces for stochastic models of biochemical reaction networks

Abstract:In stochastic models of biochemical reaction networks, the dynamics is usually represented by a Markov process which describes the evolution of the copy-numbers or molecular counts of the constituent species. It is often of biological interest to determine if this Markov process has a unique stationary distribution. This uniqueness will hold if the state-space is irreducible in the sense that all the states are reachable from each other in a finite time, with a positive probability. Finding such irreducible state-spaces is quite challenging, because the Markovian dynamics can usually access infinitely many states and the presence of conservation relations among species can constrain the dynamics in complicated ways. The aim of this paper is to develop a computational framework for finding irreducible state-spaces for reaction networks that typically arise in Systems and Synthetic Biology. Our results can help in assessing the long-term behavior of a network and also in explicitly obtaining the stationary distributions in certain cases. Our framework only relies on elementary linear algebra and linear programming, which makes it highly scalable and efficient, even for very large networks. We illustrate the wide applicability of our framework through several examples.