DYNAMICAL SENSITIVITY ANALYSES OF KINETIC MODELS IN BIOLOGY

Abstract

Mathematical models and systems analysis tools are extensively developed to understand the biological complexity. To this end, most of the existing analyses, like classical parametric sensitivity analysis, introduces static or dynamic perturbations to system states or parameters one at a time, and quantify the resulting change in system behavior. It will be shown here that the dynamical aspect of cellular regulation is not immediately apparent from the existing analysis and more importantly, that they can even mislead the modeler. To avoid the above said caveat, this dissertation presents four novel dynamic sensitivity analyses. These analyses are built on a cause-effect relationship. Based on time-varying perturbations and the resulting perturbation-effect ratio, these analyses give a dynamic species-by-species or parameter-by-parameter account of how networks function is accomplished, and also where an induced perturbation travels. In practice, the results can guide model identification and reduction of cellular systems and suggest experimentally testable hypotheses.