This dissertation focuses on the development and application of new computational methods and tools for the in silico simulation of the behavior of networks of biochemical interactions with a particular focus on synthetic gene regulatory networks.
In gene regulatory networks the protein product of one gene regulates the expression of one or more genes. With simple positive or negative feedback or feedforward regulatory relationships, biological phenotypes of astonishing complexity naturally emerge.
Gene regulatory networks are also of particular interest in the field of synthetic biology. This is a new discipline that is influenced by both engineering and biological sciences. Synthetic biology efforts focus on the construction of new gene regulatory networks that give rise to controllable phenotypic behavior. Synthetic biology is fueled by an ever-increasing body of genetic knowledge and the technologies to read and write DNA sequences quickly and inexpensively. As in any forward-engineering discipline, modeling can play a catalytic role in the rational design of synthetic gene regulatory networks.
The construction of mathematical models of gene networks can provide far more than a simple summary of experimental data. The models constructed in this dissertation are highly detailed and follow from first-principles. That is, a "reductionist" approach is followed, whereby a potentially complex network of gene expression (including all of the fundamental steps of transcription, translation, regulation, etc. of each gene) is reduced to a large series of elementary chemical reactions. These mechanistic models of gene networks are simulated using an accurate stochastic algorithm that accounts for the small size of a biological cell and the dilute nature of critical reacting species.
These tools -- quantitative, mechanistic, kinetic models simulated stochastically -- are used to study several gene network paradigms including genetic oscillators and logic gates.