Addressing biological circuit simulation accuracy: Reachability for parameter identification and initial conditions

Accurate simulation of biological networks is difficult not only due to the computational cost associated with large-scale systems simulation, but also due to the inherent limitations of mathematical models. We address two components to improve biological circuit simulation accuracy: 1) feasible initial conditions, and 2) identification of critical yet unknown model parameters. For those parameters that may not be available from experimental data, we incorporate reachability analysis to enhance our optimization/simulation framework and estimate those parameters that are capable of creating behaviors consistent with known experimental data. We apply these techniques to a biological circuit model of tryptophan blosynthesis in E. coli, and quantify the improvement in simulation accuracy when reachability analysis is used.