Metabolic P System Flux Regulation by Artificial Neural Networks

Metabolic P systems are an extension of P systems employed for modeling biochemical systems in a discrete and deterministic perspective. The generation of MP models from observed data of biochemical system dynamics is a hard problem which requires to solve several subproblems to be overcome. Among them, the flux tuners discovery aims to identify substances and parameters involved in tuning reaction fluxes. In this paper we propose a new technique for discovering flux tuners by using neural networks. This methodology, based on backpropagation with weight elimination for neural network training and on an heuristic algorithm for computing tuning indexes, has achieved encouraging results in a synthetic case study.