PANDORA

PANDORA models peptide:MHC (pMHC) complex structures to enable structural analysis for T-cell-mediated adaptive immune responses, supporting immunotherapy design and antiviral vaccine development.

Key Features:

• Generic Modelling Pipeline: Models both pMHC class I (pMHC-I) and class II (pMHC-II) complexes for broad immune interaction coverage.
• Anchor-Restrained Modelling Protocol: Identifies structural templates from an extensive database and applies anchor restraints during modelling to constrain peptide positions.
• Restrained Energy Minimization: Employs a restrained energy minimization technique to enhance modeling speed while preserving model quality.
• Competitive Accuracy and Speed: Demonstrated a median RMSD of 0.70 Å and a 93% success rate in generating top-10 models across 835 pMHC-I complexes from 78 MHC types, outperforming approaches including AlphaFold2 in comparative studies.
• Modularized Implementation: Implemented as a modular, configurable Python package.

Scientific Applications:

• T-cell recognition analysis: Provides high-quality 3D pMHC models to analyze structural determinants of T-cell receptor recognition.
• Cancer immunotherapy: Supports structural assessment of tumor-derived peptides bound to MHC for immunotherapy design.
• Antiviral vaccine development: Enables structural evaluation of viral peptides presented by MHC for vaccine candidate selection.
• Large-scale structural datasets for machine learning: Produces large-scale pMHC structural models to train and inform deep learning algorithms in immunology.

Methodology:

Searches for suitable structural templates from a comprehensive database for each query, applies anchor restraints during modelling, and refines models using restrained energy minimization.