HLA-Arena

HLA-Arena provides a customizable computational environment for structural modeling and analysis of peptide–human leukocyte antigen (HLA) complexes to support structure-based prediction of peptide–HLA binding relevant to cancer immunotherapy and neoantigen discovery.

Key Features:

• Integration with APE-Gen and DINC: Integrates APE-Gen for structure-based binding predictions and DINC for geometry prediction of peptide-HLA complexes.
• Support for MODELLER and MHCflurry: Supports MODELLER for homology modeling and MHCflurry for predicting HLA-peptide binding affinities.
• Customizable workflows: Provides customizable workflows for geometry prediction with DINC, structure-based binding prediction with APE-Gen, and large-scale virtual screening across multiple HLA alleles.
• Scalability and pipeline integration: Enables large-scale peptide screening and integration into larger computational pipelines.
• Visualization and analysis support: Includes components for structural modeling, analysis, and visualization of peptide-HLA interactions.
• Implementation: Implemented as Jupyter Notebook workflows packaged in Docker for reproducible execution.

Scientific Applications:

• Personalized cancer immunotherapy: Facilitates structure-based analysis to inform the selection of candidate peptide-HLA targets for personalized therapies.
• Neoantigen discovery: Enables structural modeling and screening of peptides to prioritize neoantigen candidates presented by specific HLA alleles.
• Vaccine development: Supports large-scale evaluation of peptide binding across HLA alleles to inform vaccine antigen selection.
• Structure-based binding prediction: Provides predicted peptide-HLA geometries and binding models to study determinants of T-cell recognition.

Methodology:

Uses DINC for peptide-HLA geometry prediction, APE-Gen for structure-based binding predictions, MODELLER for homology modeling, MHCflurry for HLA-peptide affinity prediction, and executes workflows in Jupyter Notebook packaged with Docker.