ProteinMPNN
Robust sequence design for given protein backbones
ProteinMPNN is a message-passing neural network that designs amino acid sequences folding to a specified backbone with far higher recovery and in silico success than Rosetta. It takes backbone coordinates and returns diverse, designable sequences in seconds. The standard sequence-design step after backbone generation in modern design pipelines.
ProteinMPNN is live on MolPilot — load a structure, mark what to redesign, and press Design.
At a glance
- Input
- Protein backbone (PDB)
- Output
- Designed sequences (FASTA)
- Developed by
- Baker Lab, UW
- Published
- Dauparas et al., Science 2022 · 2022
#sequence-design#inverse-folding#backbone
Related tools
More methods in Protein Design & Engineering
