Prediction of the Structures of Protein Complexes Jeffrey J. Gray, NIGMS R35 (MIRA) PROJECT SUMMARY Protein interactions are involved in nearly all biological processes in human health and diseases, and protein complex structures can reveal biological mechanisms and suggest intervention strategies. Computational modeling provides an alternative route to elucidate structures. Modeling also tests our understanding of molecular biophysics and allows us to engineer molecules based on their structures. My lab is a pioneer in developing computational methods to predict and design protein–protein interfaces and in applying those methods broadly, from immunology and cancer to infectious disease and tissue engineering. Our central focus is on protein–protein docking, that is, predicting the structure of a complex from the components (individual polypeptides or domains). A longstanding challenge in docking is correctly capturing protein conformational change, and we lead development of innovative search strategies and rapid scoring schemes to close this gap. Another focus area is the modeling and design of antibodies, directly supporting drug and diagnostic agent development and optimization. We predict high-resolution antibody structures from sequence, dock those models to antigens, and design antibodies to target specific antigens. Several protein modifications are important to consider when modeling interfaces. The emergence of powerful experimental methods for characterizing glycans has prompted us to expand our tools to model carbohydrates including docking methods and focused studies on glycotransferases. All our methods are embedded in the Rosetta software platform, which is used by tens of thousands of academic and industry scientists worldwide. The utility of our work is evidenced by the high demand for our prediction and design web server, scripting platform, and teaching materials. By leading the technical and scientific testing operations for the Rosetta Commons, my lab powers the synergies to combine Rosetta's powerful biomolecular prediction and design methods across this international collaboration.