New experimental techniques in the biosciences provide us with high-quality data allowing quantitative mathematical modeling. When fitting model parameters to experimental data, it is important to know whether all parameters can be uniquely estimated from available data. In this paper we discuss a class of models for metabolism, where the introduction of conserved moieties may cause an otherwise identifiable model to be unidentifiable. A general method for reparametrization to identifiable rate expressions is presented, and the general results are exemplified by three well-cited models for yeast metabolism.

Joint work with Milena Anguelova, Gunnar Cedersuna, Carl Johan Franzen, Mikael Johansson