Models of biochemical reaction networks, such as signaling or metabolic networks, contain a large number of components and an even larger number of parameters associated with reaction and transport processes. While the nominal parameter values are usually sourced from the literature, they may represent kinetics in a different cell type or organism than that which is modeled. As such, parameters are often estimated from the available experimental data. However, because of the limited amount of data available and the large number of parameters, regularization is needed to avoid over fitting. A tutorial of regularization techniques including parameter set selection precedes a discussion of selected state-of-the-art procedures for estimating parameters in complex biochemical networks.
Proceedings of the FOSBE 2015, Cambridge, Massachusetts (2015)