SolubleMPNN (Claude Skill)
Designs protein sequences with ProteinMPNN using soluble-optimized weights that discourage surface hydrophobics, driven as a Claude skill.
| Type | Claude Skill |
| Supplier | Baker Lab, UW Institute for Protein Design |
| Availability | GA — Claude Science research skill |
| Pricing | Free / OSS (MIT) |
| Capabilities | Read/Write — local inference; writes designed sequences (FASTA) |
| Verified | works · 2026-07-20 |
| Security | cleared · 2026-07-20 — provenance matches dauparas/LigandMPNN, MIT, no advisories |
How to install
- Claude Science — enable the built-in SolubleMPNN research skill (Anthropic-hosted; not published to the public
anthropics/life-sciencesmarketplace). - Run the model yourself — the upstream model is open source:
git clone https://github.com/dauparas/LigandMPNNFollow the repo README for environment setup and model weights.
What it does
Runs ProteinMPNN with the SolubleMPNN weight set, which penalizes solvent-exposed hydrophobic residues to improve expression and solubility of designed proteins. Local inference only. The weights ship inside the ProteinMPNN / LigandMPNN repositories.
Primary use cases: Solubility-optimized protein design, improving expression of de novo designs
Notes
Claude Science: Featured as a research skill in Anthropic’s Claude Science. Its inclusion there is an independent signal of quality and trustworthiness for life-science work.
Local (your compute). A weight variant of ProteinMPNN; bundled in the ProteinMPNN / LigandMPNN repos.
Sources
Installed this tool?
Share feedback — install path, OS, errors, workarounds. The form opens with this tool pre-selected and a link back to this page.