Use OpenBind Online
Commercially Available OpenBind No-Code Web Server
OpenBind: Accelerating Structure-Based Drug Discovery with Open Data and AI
OpenBind is a transformative, community-driven initiative building the world’s largest open dataset of protein-ligand interactions. Coordinated by Diamond Light Source in collaboration with leading global academic and industry partners—including the AlQuraishi Lab at Columbia University and the OpenFold Consortium—OpenBind bridges the data gap in structure-based drug discovery.
By combining ultra-high-throughput experimental screening with state-of-the-art open-source biomolecular AI architectures like OpenFold-3, OpenBind unlocks smarter, generalizable models to accelerate drug discovery across complex diseases.
Why OpenBind?
While revolutionary tools like AlphaFold3 and RoseTTAFold All-Atom have altered structural biology, next-generation drug discovery models remain bottlenecked by access to large, high-quality datasets of protein-ligand interactions and consistent affinity data.
OpenBind provides a comprehensive, open data foundation for the entire scientific community to push AI tools past this critical inflection point:
Massive, Open Datasets: Generates vast, FAIR-compliant protein-ligand structural datasets rather than focusing on a single company’s proprietary drug pipeline.
Smarter AI Generalization: Trains AI models to accurately predict how small molecules bind and behave across diverse therapeutic targets.
Accelerated Therapeutics: Combines automated synthesis, high-throughput crystallography, and regular blind prediction challenges to create faster, more reliable drug discovery workflows.
Key Pillars of the OpenBind Initiative
OpenBind leverages a coordinated, three-pronged framework to deliver high-value open data quickly and cost-effectively:
Pillar | Technology | Role in OpenBind |
OpenBind Circle | Ultra-High-Throughput X-Ray Crystallography | Uses Diamond Light Source's XChem and MX beamlines to collect thousands of protein:ligand complex X-ray structures every week. |
OpenBind Star | Automated Microlitre-Scale Chemistry | Enables rapid generation of crude reaction mixtures directly suitable for protein crystal soaking. |
OpenBind Triangle | Nanolitre-Scale Synthesis | Facilitates synthesis powered by extremely large chemical building block libraries. |
Advanced Structure Prediction with OpenFold-3
OpenBind integrates with OpenFold-3 (developed by the AlQuraishi Lab at Columbia University and the OpenFold Consortium), a fully open-source biomolecular structure prediction model based on DeepMind's AlphaFold3 architecture. Distributed under an Apache 2.0 license, OpenFold-3 provides:
Multi-Molar Structure Prediction: Accurately models protein, RNA, and DNA chains alongside small-molecule ligands.
High Performance: Matches AlphaFold3 performance on monomeric RNA and complex benchmarks while providing open access to training data and code.
Kernel Acceleration: Optimized for high-throughput inference using DeepSpeed4Science and cuEquivariance kernels for multi-GPU scale.
What is Tamarind Bio?
Tamarind Bio is a pioneering no-code bioinformatics platform built to democratize access to powerful computational tools for life scientists and researchers. Recognizing that many cutting-edge machine learning models are difficult to deploy, Tamarind provides an intuitive web-based environment that abstracts away high-performance computing, software dependencies, and complex command-line interfaces.
Tamarind enables biologists, chemists, and drug discovery teams to leverage published methodologies, such as OpenFold-3, AlphaFold, PepMLM, RFdiffusion, MPNN, and GROMACS, through a user-friendly graphical interface or robust API. With enterprise-grade data security verified by its Trust Center, Tamarind empowers researchers to focus on science and accelerate therapeutic development without managing cloud infrastructure.
Accelerating Drug Discovery with OpenBind on Tamarind Bio
Running OpenBind-compatible models and datasets on Tamarind Bio creates a seamless, end-to-end pipeline for structure-based drug screening and biomolecular modeling:
Data-Driven Target Screening: Screen vast chemical libraries against target proteins using ultra-high-throughput crystal structure datasets provided by OpenBind.
Open-Source Prediction: Leverage OpenFold-3 on Tamarind’s cloud environment to accurately predict protein-ligand binding interactions and structural conformations without manual local installation.
Streamlined Workflows: Rapidly iterate from target identification to fragment-based ligand screening and structural validation within a secure, high-confidence computational interface.
How to Use OpenBind on Tamarind Bio
Researchers can easily run OpenBind-aligned biomolecular models and dataset workflows on Tamarind Bio using the following step-by-step process:
Access the Platform: Log in to your account at tamarind.bio.
Select the Biomolecular Model: Choose OpenFold-3 or your designated OpenBind structural tool from the catalog of state-of-the-art models.
Input Sequence & Ligand Information: Provide the protein target amino acid sequence (or FASTA) alongside the target small molecule ligand, nucleic acids, or metadata JSON.
Configure Parameters: Specify prediction parameters, multi-query batch settings, or MSA options (e.g., using ColabFold MSA servers or customized structural templates).
Run Structure Prediction: Launch the job. Tamarind Bio manages hardware distribution across GPU clusters using efficient kernel acceleration.
Evaluate Structural Results: Download and inspect the resulting protein-ligand complex PDB/CIF structure files, confidence scores, and interaction metrics to select candidates for experimental validation.