google-deepmind/alphafold3

AlphaFold 3 inference pipeline.

What it solves

AlphaFold 3 provides a way to accurately predict the 3D structure of biomolecular interactions. This is critical for understanding how proteins, DNA, RNA, and ligands interact within biological systems.

How it works

The system consists of an inference pipeline that takes biomolecular sequences as input (via JSON files) and predicts their structure. The process is split into two main stages:

  1. Data Pipeline: A CPU-only stage that performs genetic and template searches.
  2. Inference: A GPU-accelerated stage that uses model parameters to generate the final structural predictions.

Who it’s for

Researchers in biology, biochemistry, and drug discovery who need to model the interactions between different biomolecules.

Highlights

  • Broad Biomolecular Support: Predicts interactions across a wide range of molecules including proteins, DNA, RNA, and ligands.
  • Docker-based Deployment: Provided as a Docker image for simplified installation and and execution.
  • Flexible Execution: Allows users to run the data pipeline and inference separately via command-line flags.
  • Integrated Database Support: Utilizes a wide array of mirrored reference databases (e.g., PDB, UniProt, MGnify) for structural prediction.

Related

  • Project
    aqlaboratory/openfold-3

    An open-source reproduction of AlphaFold3 for predicting the 3D structures of proteins, RNA, DNA, and small molecules.

  • Project
    lucidrains/alphafold3-pytorch

    A PyTorch implementation of AlphaFold 3 for predicting the 3D structures of biomolecular interactions, including comprehensive tools for PDB dataset curation.

  • Project
    kyegomez/Open-AF3

    An open-source PyTorch implementation of AlphaFold 3 for predicting the 3D structures of biomolecular interactions, including proteins, nucleic acids, and ligands.

  • Project
    bytedance/Protenix

    An open-source biomolecular structure prediction framework that provides high-accuracy 3D modeling of proteins and RNA, designed to be an accessible alternative to AlphaFold3.

  • Project
    sokrypton/ColabFold

    A tool that makes protein folding predictions accessible via Google Colab and local installations, integrating models like AlphaFold2 and AlphaFold3 to predict 3D protein structures.