COMPUTATIONAL PREDICTION OF MOLECULAR INTERACTIONS

Molecular interactions are of critical importance to all biological processes such as enzymatic reactions, transport, signaling, protein folding and aggregation, and macromolecular assembly. Proteins evolved to interact with other cellular and extracellular molecules in a precise and time-dependent manner. Structural analyses and modeling of such interactions is a critical step for a mechanistic description of their functions. The research in this dissertation focused on developing and applying integrative computational modeling approaches to understand the structural basis of molecular recognition and interactions, and predict modulators of these interactions. To achieve these goals, several techniques were used, such as computer simulations, physicochemical modeling, as well as graph-theoretic and machine learning techniques. More specifically, computational methods and tools were developed and applied to (1) identify ligands of clinically relevant human membrane transporters expressed in liver, kidney and intestine, (2) study molecular interactions between hemoglobin proteins in red blood cells, and (3) identify human proteins attacked by the HIV-1 protease during infection