Statistical analysis and probabilistic verification of stress-induced signalling pathways

Abstract

Recent studies reveal that dysregulation of Endoplasmic Reticulum ER stress signalling pathways is implicated in the pathogenesis of several diseases. ER is a major hub for protein synthesis, modification and sorting, and it also regulates several signalling pathways in the cell cycle progression. Disturbance of endoplasmic reticulum could induce an unfolded protein response, which is a self-protective mechanism. Graphical lasso method was first used to infer the undirected sub-networks of ER stress signalling from microarray data. Then, we construct a stochastic model to describe the crosstalk of three major signalling pathways induced by ER stress, and apply a probabilistic model checking technique to formally analyse the temporal logic properties of the model, which is written in the PRISM language. This verification technique can both qualitatively and quantitatively verify the signalling pathway model using the sequential probability ratio test and confidence interval estimation method, respectively.