Protein Design · Baker Lab
Inverse folding: given a backbone, design sequences that actually fold onto it.
1–30s
Runs on Managed GPU endpoint
Runs on managed compute. Availability and startup time depend on current capacity.Run this model
About this model
ProteinMPNN is the second half of nearly every design pipeline. Give it a backbone and it returns sequences with dramatically higher experimental success rates than Rosetta's physics-based design. Apache-2.0 licensed, fast, and reliable enough to be boring — which is the highest praise a design tool can earn.
Standardized I/O contract
Every model in the Hub speaks the same contract, which is what lets the Router and the agent call any of them without special-casing.
Inputs
Backbone (PDB)
Sequences
Sampling temperature
Outputs
Sequences with per-design score and global recovery rate.
