Parameter Estimation in Systems Biology Models by Using Extended Kalman Filter

Abstract

Models in systems biology, which reflect complex dynamic biological phenomena are most often described as ordinary differential equations (ODE). Characteristic properties of these differential equations is nonlinearity and large size (number of state variables). These models also contain large numbers of unknown parameters. So the main challenge in developing models in systems biology is estimation of numerous unknown parameters in nonlinear differential equations. There are already numerous approaches to parameter estimation in systems biology models. However, main difficulties speed of convergence and multiple minima (multiple solutions) are still obstacles in achieving solutions of sufficient efficiency. In this chapter we propose a new approach based on combination of extended Kalman filtering dynamical optimization with spline approximation of solutions to ODE, for parameter estimation in systems biology models. We present the main idea and we show comparisons to some published results.