This page brings together key documents, resources, and tools related to the OpenBind initiative, including links to data platforms and supporting materials.
Platforms
Access the Model – OpenFold-3 GitHub Repository
Access the open-source OpenFold3 codebase that OB0 is built on. Contains the training code, model weights, and training recipes, released under Apache 2.0. Take a look at the 'Documents, scripts & support resources' section of this page to access further resources.
Click to access the codeGet the Data – FatA Fragalysis Dataset
Explore the FatA fragment screening dataset hosted on Fragalysis. This provides 402 crystallographic binding events for fatty acid thioesterase A, a commercial herbicide target: 311 fragment binding events covering 128 unique fragments, plus 91 follow-on binding events covering 38 unique compounds. Data has been generated in a fragment-to-hit campaign with Syngenta.
Click to explore the datasetGet the Data – RdRp Zika and Dengue Fragalysis Dataset
Access fragment screening datasets for the RNA-dependent RNA polymerase (RdRp) domain of NS5 from zika virus (ZIKV) and dengue virus serotype 2 (DENV-2). Data includes 315 binding events for the RdRp DENV-2 and RdRp ZIKV, generated and documented by the READDI and ASAP AViDD Consortia, respectively. RdRp DENV-2 accounts for 124 (85 fragment binding events across 67 unique fragments; 39 follow-on binding events across 30 unique compounds) and RdRp ZIKV for 191 (151 fragment binding events across 99 unique fragments; 40 follow-on binding events across 33 unique compounds).
Click to explore the datasetGet the Data – Fragalysis
Fragalysis is a browser‑based platform for visualising, curating and downloading experimental and virtual fragment‑screening data. It hosts all data generated through OpenBind. Guidance on how to use the platform is available in the 'Documents' section of this page.
Click to explore FragalysisGet the data – Zenodo
This record contains the first public dataset released by the OpenBind consortium: a structure–affinity dataset for structure-based AI and computational drug discovery.
Click to explore ZenodoBenchmarks – GitHub
This record contains the first public dataset released by the OpenBind consortium: a structure–affinity dataset for structure-based AI and computational drug discovery.
Click to explore GithubExperimental protocols
OpenBind workspace to capture the protocols.io produced as a result of the consortium efforts.
Click to explore OpenBind protocolsOpenBind Data Hub
OpenBind Data Hub is a collaborative space for discussing coupled structural and assay datasets released by OpenBind. Here you can chat to the OpenBind team, troubleshoot queries, and stay updated with what's happening at OpenBind.
Click to join the OpenBind communityBlogs
Blog: OpenBind-0: Advancing Open Molecular Structure Prediction
Introducing OpenBind-0, a fully open molecular structure prediction model. This post presents the model, its performance, and new benchmark data that highlight both the promise and limitations of AI for drug discovery.
Click to read the blogBlog: OpenBind’s first release: A structure–affinity dataset for structure-based AI
Presenting OpenBind’s first structure-affinity data release for the EV A71 2A protease, alongside reference benchmarks and how the community can use and evaluate them.
Click to read the blogBlog: Inside OpenBind’s First Public Data Release: The Real Process Behind the Data Generation
Outlining the experimental and structural biology work underpinning OpenBind’s first data release, from protein production and crystallography through to model building and refinement.
Click to read the blogBlog: Affinity and Kinetics Data in the EV‑A71 2A OpenBind Release
This post documents how the first OpenBind data package was generated – from target selection and experimental design through large‑scale structure and affinity data production – highlighting the scale, complexity, and rigour behind the final dataset.
Click to read the blogDocuments, scripts & support resources
OpenBind Model Release Resources
Find the scripts, datasets, and supporting materials associated with the OpenBind model release blog post. These resources are provided to help researchers reproduce analyses, explore the model outputs, and build on the released work.
Click to access the model release resourcesFragalysis Guide
The Fragalysis app is supported by comprehensive documentation on Read the Docs, which explains how to use the platform.
Click to read the Fragalysis Guide
