Kevin Wu

Abstract

The ability to computationally generate novel yet physically foldable protein structures could lead to new biological discoveries and new treatments targeting yet incurable diseases. Despite recent advances in protein structure prediction, directly generating diverse, novel protein structures from neural networks remains difficult. In this work, we present a new diffusion-based generative model that designs protein backbone structures via a procedure that mirrors the native folding process. We describe protein backbone structure as a series of consecutive angles capturing the relative orientation of the constituent amino acid residues, and generate new structures by denoising from a random, unfolded state towards a stable folded structure. Not only does this mirror how proteins biologically twist into energetically favorable conformations, the inherent shift and rotational invariance of this representation crucially alleviates the need for complex equivariant networks. We train a denoising diffusion probabilistic model with a simple transformer backbone and demonstrate that our resulting model unconditionally generates highly realistic protein structures with complexity and structural patterns akin to those of naturally-occurring proteins. As a useful resource, we release the first open-source codebase and trained models for protein structure diffusion.

Our speaker

Kevin is currently doing his PhD in computer science at Stanford University. His research focuses on modeling biomedical datasets via biologically-motivated machine learning strategies, with the ultimate goal of better understanding biological systems and generating new molecules for potential treatments. Prior to graduate school, Kevin completed his undergraduate studies at UC Berkeley and worked at Illumina as a bioinformatics scientist.

To become a member of the Rough Path Interest Group, register here for free.