MHCcluster, a method for functional clustering of MHC molecules

Area of Science:

• Immunology
• Bioinformatics
• Computational Biology

Background:

• Major Histocompatibility Complex (MHC) molecules present peptides to T cells, crucial for immune response.
• The human leukocyte antigen (HLA) system is highly polymorphic, with most alleles uncharacterized for binding specificity.
• Characterizing MHC binding specificities is essential for understanding T cell immunity across species.

Purpose of the Study:

• To introduce MHCcluster, a novel tool for functionally clustering MHC molecules based on predicted peptide binding.
• To demonstrate the limitations of sequence-based clustering for MHC molecules.
• To analyze functional diversity of MHC class I molecules in humans and chimpanzees.

Main Methods:

• Development of the MHCcluster web tool for flexible analysis of user-selected MHC molecules.
• Utilizing predicted peptide binding specificities for functional clustering of MHC variants.
• Employing heat maps, tree-based visualizations, and a dynamic TreeViewer for data representation.

Main Results:

• MHCcluster successfully identifies functional relationships between MHC molecules, outperforming sequence-based clustering.
• Clustering of prevalent HLA-A and HLA-B alleles revealed 12 major specificity groups (supertypes), with some alleles not fitting existing classifications.
• Chimpanzee MHC class I molecules exhibit reduced functional diversity compared to human HLA class I molecules.

Conclusions:

• Functional clustering based on predicted binding specificity is superior to sequence-based methods for MHC analysis.
• MHCcluster provides valuable insights into MHC polymorphism and functional diversity, aiding T cell immunology research.
• The tool facilitates the identification of MHC supertypes and unclassified alleles, advancing our understanding of immune responses.