Algorithms for Discriminating Between Biochemical Reaction Network Models: Towards Systematic Experimental Design

We develop a methodology for distinguishing between biochemical reaction network models for biological systems modeled by nonlinear ordinary differential equations. We show how to choose the 'best' initial condition direction (the dependent variables, the state) so as to maximize the discrepancy between the outputs of possible network models. This information could then be used to design new experiments with the hope that some of the networks would be invalidated from experimental data. The methodology is applied to discriminate between models of the network underlying chemoattractant-induced movement of
Dictyostelium
cells.