An improved stochastic modelling framework for biological networks

Abstract

It has become very clear that stochasticity in biology is a rule rather than exception. Gillespie stochastic simulation algorithm (GSSA) (direct method) is the first algorithm proposed to model stochasticity in biochemical systems. However, the computational intractability of direct method has been identified as the main challenge for using it to model large biochemical systems. In this paper, a novel variant of the GSSA is proposed to address computational intractability of the direct method. The direct method is combined with a Mapping Reduction Method (MRM) to target a single run of the direct method to be accelerated by advancing the system through several reactions at each time step to replace the single reaction in GSSA. MRM is a framework for mimicking parallel processes occurring in large systems using a large number of threads that work together and seen as a single system. It is used for parallel problems to be processed across large datasets using a large number of nodes working together as a single system. Link between GSk3 and p53 in Alzheimer's disease (AD) is modelled using the proposed method and tested and validated by comparing it with the direct method.