aurekaresearch/OpenDDE

An Open-source Drug Discovery Engine

What it solves

OpenDDE is designed to accelerate drug discovery by providing a scalable engine for biomolecular structure prediction, design, and optimization. It addresses the complexity of predicting how different biomolecules (proteins, DNA, RNA, and ligands) interact and fold together, a process known as co-folding.

How it works

It is an all-atom biomolecular foundation model that utilizes co-folding to predict structures. The system supports various input types, including protein chains, DNA, RNA, ligands, and ions. To improve accuracy, it can integrate external data through protein Multiple Sequence Alignments (MSA), template searches, and RNA MSA preprocessing. For larger inputs, it features a "Fold-CP" inference mode that allows the model to run across multiple GPUs using context-parallelism.

Who it’s for

This tool is primarily for researchers in drug discovery, structural biology, and computational chemistry who need to predict the 3D structures of complex biomolecular assemblies.

Highlights

  • All-atom foundation model: Predicts structures for proteins, nucleic acids, and small molecules.
  • Co-folding capabilities: Specifically designed to handle the interaction and folding of multiple biomolecular components.
  • Multi-GPU scaling: Supports Fold-CP for distributed inference on larger molecular inputs.
  • Specialized checkpoints: Includes a general-purpose model and a version specifically tuned for antibody-antigen interactions.

Related

  • Project
    aqlaboratory/openfold-3

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

  • Project
    google-deepmind/alphafold3

    An implementation of the AlphaFold 3 inference pipeline for the accurate 3D structure prediction of biomolecular interactions.

  • 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.

  • Project
    OTeam-AI4S/ODesign

    An all-atom generative world model for biomolecular interaction design that generates proteins, ligands, and nucleic acids tailored to bind to specific targets.

  • 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.