Living biological systems are complex systems which contain various components and interactions. One major challenge is to understand, at the system level, biological systems that are composed of components revealed by molecular biology. Systems biology is a rapidly growing interdisciplinary area in which modeling, system analysis and design are fundamental. Various complex phenomena emerge in living cells stimulated from interior or exterior stimulations. Periodic changes of biochemical and biophysical quantities are a ubiquitous phenomenon in living systems. Model-based analysis can help to understand the biophysical mechanisms underlying the oscillatory dynamics, which is important in revealing the nature of living systems. In this dissertation, oscillatory biological networks are investigated by model-based bifurcation analysis, sensitivity analysis and robustness analysis. The main contributions are summarized as follows. ⑴ Based on literature study and analysis using theory of system dynamics, the biophysical mechanisms of cellular oscillations are investigated. Two biological oscillators, the Drosophila circadian rhythm gene networks and NF-B signal transduction networks, are examined in detail. The study shows that negative feedback and effective time delay are the sources to produce sustained oscillations. ⑵ Following dynamics system theory, bifurcation analysis is employed to analyze NF-B signal transduction networks and Drosophila circadian rhythm. The numerical results for NF-B signal transduction networks indicate that this system can produce limit cycle oscillations under certain conditions, which supports the observation of damped oscillations in experiments and are in favor of the possibility of sustained oscillations in vivo. For the Drosophila circadian rhythm, bifurcation analysis reveals that, even for a typical oscillator, sustained oscillations can only occur within limited parameter ranges. ⑶ A modified method based on singular value decomposition is proposed to obtain period sensitivity through the study of quasi-periodic accumulation in state sensitivity time series. This method has good convergence and is easy to interpret and easy to implement. The period sensitivity can be calculated systematically without introducing further perturbations to determine direction of each period sensitivity. ⑷ On investigating the constituents of raw state sensitivity, and following the concept of phase definition,...
修改评论