Massively Parallel Synthetic Investigation of Protein Folding and Binding
Seminar | February 7 | 12-1 p.m. | 106 Stanley Hall
Gabriel Rocklin, University of Washington
Computationally designed miniproteins (~40 amino acids in length) have exciting potential as therapeutics and are also ideal model systems for studying protein biophysics. We recently introduced massively parallel methods for designing, expressing, and experimentally testing tens of thousands of entirely new proteins, leading to an avalanche of new protein structures. Assaying these "de novo" designed proteins for folding enabled us to discover over 2,500 well-folded miniproteins in four different folds, and assaying designed libraries for binding identified hundreds of active binders to the design targets Influenza Hemagglutinin and Botulism neurotoxin B. The datasets from these large-scale experiments revealed fundamental principles of folding and binding and provided clear direction on how to improve the design process, leading to higher success rates and new structures unlike those found in nature. These feedback cycles promise to transform de novo protein design into a data-driven science and to unlock the potential of de novo design for therapeutic development.