PME Immuno Special Seminar - Dr. Jason Zhang

AI-generated de novo Design of Proteins to Track and Drug Cancer Signaling In this work, I present advances in developing and using AI-driven de novo protein design to generate tools for probing and manipulating cellular signaling, focusing on Ras proteins. I describe the use of computational protein design to create genetically encodable biosensors, such as those that are capable of measuring endogenous Ras activity and profiling its signaling environment with high spatiotemporal resolution. Additionally, I present new methodologies to overcome the challenges of targeting unstructured, highly charged protein regions such as the C-terminus of Ras to create highly specific binders for Ras isoforms and peptide-MHC cancer neo-epitopes. These tools reveal new insights into the compartmentalized Ras signaling, mechanisms driving drug resistance, and development of new cancer immunotherapies. By integrating deep learning and protein design, this approach offers a framework for studying complex biochemical networks and developing novel strategies for cancer biology, diagnostics, and therapeutics.