Computational methods are becoming increasingly important for studying complex biochemical systems. As computers have become faster and more powerful, it is possible to perform calculations of larger and more complex systems than has been possible in the past. As these systems increase in complexity, it becomes increasingly important to develop methods that efficiently utilize the available computational power, and use new tools and methods to search for meaningful results. The work presented in this dissertation uses computational methods for studies in two major areas of computer aided chemistry research - diffusion and drug discovery. Diffusion is often a rate determining step in many biochemical processes, therefore being able to study multi-step diffusion reactions is important for learning about how diffusion regulates reactions in complex biological systems. The diffusion work presented in the first chapter of this dissertation utilizes simplified spherical models to perform diffusion studies of a two-step reaction model. The results give insight into how the relative locations of reaction targets are important in determining the success of a multi-step diffusion reaction. The second chapter of this dissertation proposes a new hybrid method for diffusion studies that capitalizes on the benefits of two popular calculation methods, Brownian dynamics simulations and finite element methods. The hybrid method utilizes the speed of finite element method calculations for a majority of the diffusion system, but allows for atomistic detail of the target and diffusing particles in the region of interest near the site of the reaction. Basic one- dimensional test cases are presented to demonstrate that the method is accurate and viable for future development. Though the systems presented in both diffusion studies are basic test cases, the results provide a useful basis for future studies of more complicated biological systems. Another major area of computational chemistry research is that of computer aided drug discovery. The third chapter of this dissertation uses molecular dynamics simulations and molecular docking tools to search for potential drug compounds to inhibit the bacterial enzyme dihydropteroate synthase. Visualization tools are utilized to identify structures of interest, and several hundred compounds are identified for further testing

Nature often colocalizes successive steps in a metabolic pathway. Such organization is predicted to increase the effective concentration of pathway intermediates near their recipient active sites and to enhance catalytic efficiency. Here, the pathway of a two-step reaction is modeled using a simple spherical approximation for the enzymes and substrate particles. Brownian dynamics are used to simulate the trajectory of a substrate particle as it diffuses between the active site zones of two different enzyme spheres. The results approximate distances for the most effective reaction pathways, indicating that the most effective reaction pathway is one in which the active sites are closely aligned. However, when the active sites are too close, the ability of the substrate to react with the first enzyme was hindered, suggesting that even the most efficient orientations can be improved for a system that is allowed to rotate or change orientation to optimize the likelihood of reaction at both sites.