Abstract

The recent developments in biological experiments have awarded the research community with valuable information, which describe finely regulated systems that govern the cell dynamics. One of the greatest challenges, however, remains to represent this extensive amount of knowledge in a proper way that can be used in simulations, and validated automatically, in order to understand the dynamics and ultimately achieve a desired behaviour for the system (cell) under control.
Many tools and techniques have been proposed in the literature to address this important problem. In this research, the use of Petri nets for knowledge representation is investigated. The initial focus of this research is then to introduce a concept of consistency between Petri nets obtained from various knowledge sources. Two algorithms are provided to construct Petri net models for cell dynamics using data available in public domain biological database. The first algorithm generates a low-level model capturing protein- protein interactions and the second, produces a high-level model which describes pathway sequences and is considerably easier to analyze. Appropriate tests are developed to study consistency of such models.
In the context of biological systems, diseases that alter cell dynamics, such as cancer, can be regarded as faults in the system, and disease diagnosis and treatment will correspond to fault detection and control. In this research a framework has been proposed for sensitivity analysis in Petri net representation of biological systems. Efficient tools and procedures are developed to achieve sensitivity analysis. It is demonstrated using actual biological system models, that the results of such analysis can be used as a basis of drug discovery.