> ## Documentation Index
> Fetch the complete documentation index at: https://docs.adaptyvbio.com/llms.txt
> Use this file to discover all available pages before exploring further.

# ProteinMPNN

> Backbone-conditioned sequence design using message passing neural networks

<Info>
  ProteinMPNN designs sequences for a fixed backbone by maximizing sequence-backbone compatibility via message passing. Widely used for inverse folding and binder design.
</Info>

<CardGroup cols={2}>
  <Card title="Type" icon="layer-group">Inverse folding (sequence design)</Card>
  <Card title="Input" icon="download">Backbone coordinates (PDB)</Card>
  <Card title="Output" icon="upload">Sequences consistent with backbone</Card>
  <Card title="Origin" icon="building">Baker Lab</Card>
</CardGroup>

## Resources

* GitHub: [dauparas/ProteinMPNN](https://github.com/dauparas/ProteinMPNN)
* Paper: [Robust deep learning–based protein sequence design using ProteinMPNN](https://science.org/doi/10.1126/science.add2187)

<Tip>
  Pair with structure predictors (e.g., ESMFold/AlphaFold2) to validate designed sequences.
</Tip>