Overview
Helper scripts parse PDB inputs and specify chains, fixed residues and sequence constraints. It is a structure-conditioned sequence model, not a backbone generator.
Limitations
A suitable backbone is required. The observed default-branch commit is old; ongoing active maintenance is not established. Designs need further evaluation.
Key Features
- Backbone-conditioned sequence design
- Chain and residue constraints
Use Cases
- Generate sequences for candidate backbones
- Study designs with fixed residues
How to Use
Prepare the documented Python dependencies, parse an example PDB and run protein_mpnn_run.py with the chosen model, chains and constraints.