A geometric analysis of fast-slow models for stochastic gene expression (Q907114)

In this paper, the authors study gene expression in prokaryotic and eukaryotic organisms that complicate the modelling of gene regulatory networks due to the high stochasticity of the process. The analysis relies crucially on the probability-generating function \(F\) that is induced by the propagator probabilities. It is shown that \(F\) solves the linear first-order partial differential equation of the form \[ \frac{\partial F}{\partial\tau}+\gamma[u-b(1+u)v]\frac{\partial F}{\partial u}+v\frac{\partial F}{\partial v}=auF, \] where \(F\) must satisfy \[ F(u,v,\tau=0,\gamma^{-1})=(1+u)^{m_0}(1+v)^{n_0},\;m_0,n_0\in\{0,1,2,\dots\}. \] Under the assumption that \(\gamma\gg1,\) which means that the degradation rate of mRNA is much larger than that protein, they introduced \(\epsilon=\gamma^{-1}\) as a small perturbation parameter. Using the method of characteristics the partial differential equation above is rewritten into the system of ordinary differential equations \[ \begin{aligned} \frac{d\tau}{dr}&=1, \\ \epsilon\frac{du}{dr}&=u-b(1+u)v, \\ \frac{dv}{dr}&=v, \\ \frac{dF}{dr}&=auF \\ \end{aligned} \] and this system is analyzed in the framework of geometric theory of singular perturbations. The first-order asymptotics (in \(\epsilon\)) of the probability distributions subsequently is derived. Finally, the results are numerically verified and their practical applicability is discussed.