SCEptRe

SCEptRe generates regularly updated receptor-specific structural benchmark datasets by extracting and clustering 3D antigen–receptor complexes from the Immune Epitope Database (IEDB) to support development and validation of epitope prediction methods.

Key Features:

• Weekly extraction of 3D structures: Extracts weekly updated 3D structural complexes from the Immune Epitope Database (IEDB).
• Supported complex types: Includes antibody–antigen, T-cell receptor–peptide–MHC (TCR-pMHC), and MHC-ligand complexes.
• Clustering and dataset generation: Clusters complexes by antigen, receptor, and epitope features to produce benchmark datasets.
• Customizable filtering: Enables selection of structural quality thresholds and clustering parameters such as resolution, R-free factor, and antigen or epitope sequence identity.
• Receptor annotation: Provides annotated complementarity-determining region (CDR) sequences and V(D)J gene segment information for immune receptors.
• Non-redundant benchmarks: Produces comprehensive non-redundant benchmark datasets for evaluating epitope prediction methods.

Scientific Applications:

• Epitope prediction development: Provides receptor-specific benchmark datasets for development and validation of epitope prediction algorithms.
• Algorithm benchmarking: Enables performance evaluation of new-generation epitope prediction methods using structurally resolved datasets.
• Vaccine design and immunogen selection: Supports vaccine design and immunogen selection by supplying structurally characterized antigen–receptor interaction data.
• Diagnostics and therapeutics: Aids disease diagnostics and therapeutic strategy development by providing receptor-specific epitope information.

Methodology:

Extracts weekly updated 3D structural complexes from the Immune Epitope Database (IEDB); identifies antibody–antigen, TCR-pMHC, and MHC-ligand complexes; clusters complexes based on antigen, receptor, and epitope features; applies structural quality and clustering filters (resolution, R-free factor, antigen/epitope sequence identity); annotates receptors with CDR sequences and V(D)J gene segments; and outputs non-redundant benchmark datasets.