Identifiablity Analysis and Improved Parameter Estimation of a Human Blood Glucose Control System Model

摘要

Quantitative dynamical mathematical models are very useful in the thorough understanding and possible targeted manipulation of biological processes. However, determining the model parameters from available data is often a challenging task for such models, typically given in the form of nonlinear ordinary differential equations. Global structural identifiability of parameterized ODE models means that there is (at least a theoretical) possibility to uniquely determine system parameters from appropriate measurement data [2]. The aim of this paper is to study structural identifiability for a published molecular level model of human blood glucose control, to achieve improvement in model fit compared to published results, and thus to obtain a model that will be suitable to examine the effect of natural and artificial feedbacks.