AlphaFold2 (Claude Skill)

Predicts 3D protein and complex structures with AlphaFold2, driven as a Claude skill over local or vendor-hosted inference.


Type	Claude Skill
Supplier	Google DeepMind
Availability	GA — Claude Science research skill
Pricing	Free / OSS (Apache-2.0 code; AlphaFold2 parameters CC BY 4.0)
Capabilities	Read/Write — Claude runs the model and writes predicted structures (PDB/mmCIF) plus per-residue pLDDT/PAE confidence
Verified	works · 2026-07-29
Security	cleared · 2026-07-29 — provenance matches google-deepmind, Apache-2.0, maintained, no OSV advisories

How to install

Claude Science — enable the built-in AlphaFold2 research skill (Anthropic-hosted; not published to the public anthropics/life-sciences marketplace).
Run the model yourself — the upstream model is open source:
```
git clone https://github.com/google-deepmind/alphafold
```
Follow the repo README for environment setup and model weights.

What it does

Runs AlphaFold2 to predict monomer and multimer structures from sequence. Builds multiple-sequence alignments locally or via the ColabFold MSA server (--use_msa_server, api.colabfold.com), or retrieves precomputed models from the AlphaFold DB API (alphafold.ebi.ac.uk/api). Emits ranked structures with pLDDT and PAE confidence.

Primary use cases: Structure prediction for a target list, complex/interface modelling, pre-docking model preparation

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.

Runs locally (GPU recommended) or via a vendor-hosted API. For retrieval of already-deposited predictions (no compute) see the AlphaFold MCP Server; this skill does de novo prediction.

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.