Comparison of automated parameter estimation methods for neuronal signaling networks

This work is a suitability study of the different optimization methods for automated parameter estimation (fitting) in the context of neuronal signaling networks. The Gepasi simulation software is used in this study since it provides a relatively good variety of optimization methods. All the available methods are used to estimate the values of reaction rate coefficients for protein kinase C signaling pathway, an important neuronal signal transduction pathway. The results show that stochastic optimization methods generally outperform the deterministic ones. Based on the results of this work, we conclude that the so-called hybrid methods should be developed in the future.