LigandMPNN vs ProteinMPNN

LigandMPNN: Protein sequence design that explicitly models small molecules, nucleotides and metals, and also outputs side-chain conformations. ProteinMPNN: Deep-learning inverse folding: designs sequences for given backbones.

Facts below come from each project's own website, repository and paper. We don't publish benchmark claims here; see the papers for evaluations. Shared tags: inverse-folding, design.

FieldLigandMPNNProteinMPNN
TaskSequence design (inverse folding) with ligand, nucleotide and metal context; side-chain packingInverse folding (sequence design)
ModalitiesProtein with small molecules, nucleotides, metalsProtein, single or multiple chains
MSA required——
Open weightsYes (MIT)Yes (in repo)
GPU need——
LicenseMIT (code and model parameters MIT per README)MIT
PricingOpen sourceOpen source
Released2023-12-232022-06-04
PaperLinkLink
GitHubGitHubGitHub
GitHub stars6451,878
InterfacePackagePackage

License and commercial use

FieldLigandMPNNProteinMPNN
Code licenseMITMIT
Weights licensecode and model parameters MIT per README—
Commercial usePermissive (commercial OK)Permissive (commercial OK)
AnswerYes. The recorded license is permissive, so closed-source commercial use is allowed if you keep the notices. DetailsYes. The recorded license is permissive, so closed-source commercial use is allowed if you keep the notices. Details

Add more Protein structure & design tools to this comparison →

More Protein structure & design comparisons

← All Protein structure & design tools