This dissertation uses modeling and simulation in order to better understand the immune response. Numerous in vivo and in vitro studies have elucidated the basic cellular mechanisms that underlie many aspects of immune response dynamics. However, in most cases it is not well understood how these mechanisms fit together. This is particularly true for the germinal center reaction (an important component of the immune response). Although many models of the germinal center have been proposed, most only attempt to explain the average-case, qualitative behavior of the system. This dissertation demonstrates that such a methodology can be misleading and presents a number of contributions toward the development of discrete/stochastic, quantitative models of the germinal center reaction.

The work presented in this dissertation is a fusion of computer science and theoretical immunology. Formulas for estimating response-specific model parameters, algorithms for developing discrete/stochastic simulations, and quantitative constraints are combined to provide a framework for asking precise questions about how well germinal center models can explain experimental observations. This framework is used to highlight significant weaknesses in well-known models of the germinal center reaction. To address these problems, new models are proposed and validated. All of these models are used to make specific predictions that can be tested by in vivo experiments to obtain further validation and drive the development of improved models.