Parameter identification for nonlinear biological phenomena modeled by S-systems

Abstract

For computational modeling of biological systems, one of the major challenges is the identification of the model parameters. It is very beneficial to use easily obtained measurements and estimate variables and/or parameters in such systems. For instance, time-series dynamic genomic data can be used to develop models representing dynamic genetic regulatory networks. These models can be used to design intervention strategies such as understanding the biological system behavior and curing major illnesses. The study shown in this paper focuses on the parameter identification of biological phenomena modeled by S-systems using Particle Filter (PF). While the nonlinear observed system is assumed to progress according to a probabilistic state space model, the results show that the PF has good convergence properties. It is concluded that the good convergence is due to PF's ability to deal with highly nonlinear process models.