rfdiffusion

What problem does it solve?

RFdiffusion backbone generation provides a method to design novel protein backbones for binders, scaffolds, and symmetric assemblies, streamlining de novo structure creation.

Core Features & Use Cases

• De novo backbone generation: Create new protein backbones using RFdiffusion with configurable contigs and hotspots.
• Motif scaffolding and symmetry: Support motif insertion and multi-chain symmetric designs.
• Integrations: Pair with ProteinMPNN for sequence design and Alphafold/Chai for structure validation; use Protein-QC for QC filtering.

Quick Start

Run RFdiffusion to generate backbones for a target PDB using a contigmap and hotspot specification, for example: modal run modal_rfdiffusion.py --pdb target.pdb --contigs "A1-150/0 70-100" --hotspot "A45,A67,A89" --num-designs 100 (Alternatively, run locally with python run_inference.py ...)