Protein Design with Generative Diffusion Models

Explore the cutting-edge applications of generative diffusion models in protein design through this comprehensive lecture featuring presentations by Sarah Alamdari and Ava Amini. Begin with an introduction to diffusion models, understanding their theoretical foundations and formulation as state-of-the-art generative deep learning approaches. Discover how these models are revolutionizing protein sequence and structure generation, rivaling the capabilities seen in image and text generation. Delve into the biophysics-inspired approach to protein structure design, mirroring the native protein folding process. Examine the innovative EvoDiff framework, which combines evolutionary-scale data with diffusion models for controllable protein design in sequence space. Gain insights into the potential of these modeling frameworks to enable new capabilities in protein engineering, paving the way for programmable, functional design in applications spanning pharmaceuticals, molecular tools, and synthetic biology.