# dauparas/ProteinMPNN

Code for the ProteinMPNN paper

Repository: https://github.com/dauparas/ProteinMPNN
Canonical: https://ross.abutalabs.com/products/proteinmpnn
Language: Jupyter Notebook
License: MIT
License Family: permissive
Last push: 2024-08-14T02:57:30+00:00

## Health v2 (maintenance only)
Score: 23/100 (v2, computed 2026-09-03T02:39:23.370411+00:00)
- activity 0, release rhythm 8, longevity 100
- inputs: {"age_days": 1561, "days_push": 749, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1834, forks 491 (observed 2026-08-28T04:05:42.464139+00:00)

## What it is
ProteinMPNN is a PyTorch-based tool that designs amino acid sequences for given protein backbone structures using a message-passing neural network. It includes pretrained full-backbone and CA-only model weights, helper scripts for PDB parsing and residue fixing, and Colab notebooks.

## Use cases
- design protein sequences for a fixed backbone structure
- generate novel sequences that fold into a target PDB backbone
- score an existing sequence against a protein backbone
- compute per-residue sequence probabilities from a structure
- fix specific residues and design the rest of a protein
- add amino acid biases or tie residues during design
- retrain the sequence design model on custom data

## When to choose
- you need fast, accurate inverse protein folding (sequence design from backbone)
- you want a well-cited, widely used tool with pretrained weights and MIT licensing
- you need fine control over which residues to fix, bias, or tie

## When to avoid
- you need full protein structure prediction rather than sequence design
- you lack a GPU or PyTorch environment and need a no-setup web tool only
- your task is unrelated to protein engineering or structural biology

## Facets
- artifact type: library
- maturity: stable
- function: machine-learning, deep-learning, llm-inference
- domain: bioinformatics, machine-learning, deep-learning
- platform: python, windows
- tags: protein-design, proteinmpnn, sequence-design, pytorch, computational-biology, structure-prediction, pdb, linux, macos

## Member repositories
- dauparas/ProteinMPNN (main) score 23

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:05:42.464139+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T03:18:35.631408+00:00, confidence not recorded.
  - readme: https://github.com/dauparas/ProteinMPNN (fetched 2026-08-28T04:05:42.464139+00:00, sha 988074603c85)
- Data as of 2026-08-30T08:39:29.467469+00:00.