Baker, DavidKing, Chris2014-04-302014-04-302014-04-302014King_washington_0250E_12847.pdfhttp://hdl.handle.net/1773/25382Thesis (Ph.D.)--University of Washington, 2014Proteins possess huge potential as therapeutic agents for the control and modulation of human physiology. Protein interactions regulate most physiological processes, mediating the connection between atomic-­scale self­-assembly and macroscale health and disease. Natural proteins often display exquisite specificity and high affinity for molecular targets while avoiding detection and elimination by the host immune system. The design of synthetic, non­-natural proteins to bind these molecular targets presents the opportunity to suppress, regulate, or enhance the cellular control processes underlying the physiology of both normal and disease states. Here, we demonstrate the development, application, and testing of computational protein design algorithms to predict protein-­binding specificity, model the energetics of designed protein interactions, and reduce the immunogenicity of protein therapeutics.application/pdfen-USCopyright is held by the individual authors.biotherapeutics; deimmunization; machine learning; molecular modeling; protein designBiochemistryBioinformaticsBiophysicsbiological chemistryThesis