RFdiffusion

De novo design of proteins, binders and motifs by diffusion

RFdiffusion denoises protein backbones with a RoseTTAFold-derived network to generate novel monomers, oligomers, binders, enzyme motifs and symmetric assemblies from scratch. Conditioning inputs define scaffolding problems such as motif grafting or target-binding constraints. Output backbones feed directly into ProteinMPNN for sequence design.

RFdiffusion is being onboarded — request access and be first in line.