Broadly applicable and accurate protein design by integrating structure prediction networks and diffusion generative models

By Joseph L. Watson, David Juergens, N. Bennett, Brian L. Trippe, Jason Yim, Helen E. Eisenach, Woody Ahern, A. J. Borst, R. Ragotte, L. Milles and 17 more. 2022

200 citations recorded by the source corpus.

There has been considerable recent progress in designing new proteins using deep learning methods1–9. Despite this progress, a general deep learning framework for protein design that enables solution of a wide range of design challenges, including de novo binder design and design of higher order symmetric architectures, has yet to be described. Diffusion models10,11 have had considerable success in image and language generative modeling but limited success when applied to protein modeling, likely due to the complexity of protein backbone geometry and sequence-structure relationships. Here we show that by fine tuning the RoseTTAFold structure prediction network on protein structure denoising tasks, we obtain a generative model of protein backbones that achieves outstanding performance on unconditional and topology-constrained protein monomer design, protein binder design, symmetric oligomer design, enzyme active site scaffolding, and symmetric motif scaffolding for therapeutic and metal-binding protein design. We demonstrate the power and generality of the method, called RoseTTAFold Diffusion (RFdiffusion), by experimentally characterizing the structures and functions of hundreds of new designs. In a manner analogous to networks which produce images from user-specified inputs, RFdiffusion enables the design of diverse, complex, functional proteins from simple molecular specifications.

Subfields: Computational Theory and Mathematics, Molecular Biology

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computational Theory and Mathematics

In Molecular Biology

Citations in the corpus

Improving and generalizing flow-based generative models with minibatch optimal transport (2023, unknown): “Diffusion models owe their success in part to their simple regression training objective, which does not require simulating the SDE during training.”
Practical and Asymptotically Exact Conditional Sampling in Diffusion Models (2023, unknown): “We next evaluated TDS on a benchmark set of 24 motif-scaffolding problems and compare its success rates to the previous state of the art, RFdiffusion (Watson et al., 2022).”
Equivariant 3D-conditional diffusion model for molecular linker design (2022, unknown): “Dataset processing was done in Python (v.3.10.5) using RDKit (v.2022.03.2) for generating molecular conformers and splitting them in fragments and linkers, scikit-learn (v.1.0.1) for splitting datasets and BioPython (v.1.79) for processing protein structures.”
Protein structure generation via folding diffusion (2022, background): “Of notable success among these methods is RFDiffusion 18 , which not only presents a myriad of conditional generation applications that can design proteins binding speci ﬁ c targets, but also performs thorough experimental validation of computationally generated proteins.”
Improving de novo protein binder design with deep learning (2022, unknown): “An updated version of RoseTTAFold (RF2 22,23 ) was also found to predict all monomer structures with binder C ɑ accuracy between 0.2 Å − 0.8Å, except for TrkA which was predicted with 1.8 Å accuracy (Supplementary Fig.”
Efficient and accurate prediction of protein structure using RoseTTAFold2 (2023, background): “We have found that RF can be readily extended to protein design by fine-tuning for inpainting 15 and diffusion tasks 16 .”