Unlock Antibody Dynamics: ABB4-STEROIDS is Now Available on Vecura

What is ABB4-STEROIDS?

ABB4-STEROIDS is a generative flow-matching model designed to predict conformational ensembles of antibodies directly from paired VH and VL sequences. By utilizing flow matching on SE(3), the model moves beyond static structural prediction to capture the inherent flexibility of CDR loops in solution. It is especially useful for researchers who need to characterize antibody dynamics, calculate ensemble-averaged binding energies, or select candidates with structural diversity for experimental validation.

What can users do with ABB4-STEROIDS on Vecura?

With ABB4-STEROIDS on Vecura, users can:

• Generate Conformational Ensembles: Quickly produce a range of 3D antibody structures from paired sequence inputs, reflecting true conformational diversity.
• Normalize via IMGT Numbering: Automatically renumber predicted PDB files using the Anarcii model, ensuring consistent CDR loop identification.
• Analyze Structural Variability: Calculate pairwise CDR backbone RMSD and per-residue RMSF to quantify loop flexibility.
• Streamline Workflows: Skip the complex technical setup of PyTorch environments and distributed runners, accessing the full power of the model through an integrated pipeline.

What the output means

The output provides a comprehensive set of PDB files representing a conformational ensemble, along with analytical CSV files detailing CDR-specific RMSD and RMSF metrics. These metrics quantify how structurally variable each CDR loop is across the ensemble.

This output should be used to support scientific decision making, such as prioritizing drug candidates or informing binding studies. It does not replace experimental validation, but rather provides a high-fidelity starting point for structural analysis.

Why this matters

Traditional antibody structure prediction often relies on static models, which fail to account for the dynamic behavior essential to antibody-antigen recognition. By generating an ensemble of structures, ABB4-STEROIDS allows scientists to explore the "breathing" of CDR loops, leading to more accurate insights into binding dynamics and stability that static models simply cannot provide.

This approach is transformative for fields like therapeutic antibody design, where understanding how loop flexibility affects binding affinity can be the difference between a successful drug candidate and a failed one. By democratizing access to this tool, researchers can now conduct advanced ensemble-based analyses without needing extensive computational infrastructure.

• Developed by: Oxford Protein Informatics Group (OPIG)
• Source: Official GitHub Repository
• Reference: Original paper (bioRxiv preprint)