Controllable protein design with particle-based Feynman-Kac steering

TL;DR

This paper introduces a Feynman-Kac framework to steer diffusion-based protein design models, enabling control over properties like interface energetics and binder designability with significant improvements.

Contribution

It develops guiding potentials using ProteinMPNN and structural relaxation to steer diffusion models toward desired protein properties, a novel model-independent approach.

Findings

Improved predicted interface energetics through steering.

Increased binder designability by 89.5%.

Demonstrated effective steering toward arbitrary, non-differentiable objectives.

Abstract

Proteins underpin most biological function, and the ability to design them with tailored structures and properties is central to advances in biotechnology. Diffusion-based generative models have emerged as powerful tools for protein design, but steering them toward proteins with specified properties remains challenging. The Feynman-Kac (FK) framework provides a principled way to guide diffusion models using user-defined rewards. In this paper, we enable FK-based steering of RFdiffusion through the development of guiding potentials that leverage ProteinMPNN and structural relaxation to guide the diffusion process towards desired properties. We show that steering can be used to consistently improve predicted interface energetics and increase binder designability by $89.5\%$. Together, these results establish that diffusion-based protein design can be effectively steered toward arbitrary, non-differentiable objectives, providing a model-independent framework for controllable protein generation.