SaProt

Creates protein embeddings that use structure as well as sequence.

Field
Biology & genomics
Method
Foundation model
Data it takes
sequence, structure
License
MIT AND GPL-3.0 — restricted use

SaProt encodes each residue using both its amino acid and its local structural context, via the 3Di alphabet from foldseek. Where ESM C uses sequence alone, SaProt also uses structure.

This benefits tasks where function depends more on shape than on sequence similarity.

The container bundles foldseek, so the license is part MIT and part GPL-3.0.

Upstream project →

Catalog entry last checked 2026-08-17. All models →

Run SaProt on your data

The lab runs this on its own compute, in a container, with the inputs and parameters recorded alongside the result. Initial scoping conversations are free.

Contact the Lab