CAMAP

CAMAP predicts MHC-I associated peptide (MAP) presentation by modeling how codon arrangement in mRNA sequences flanking MAP-coding codons (MCCs) affects peptide generation for CD8 T cell–mediated immune recognition.

Key Features:

• Artificial neural network: Employs an artificial neural network model to predict MAP presentation from mRNA flanking sequences.
• Flanking codon-based prediction: Uses only codon sequences flanking MCCs and deliberately excludes the MCCs themselves from input.
• Codon arrangement specificity: Shows improved predictive accuracy with original codon sequences compared to shuffled sequences that preserve amino acid usage.
• Independence from expression and binding: Predictive signals are reported to be independent of mRNA expression levels and MHC-I binding affinity.
• Score integration: CAMAP scores can be integrated with MAP ligand scores and transcript expression levels to improve MAP prediction.
• Experimental support: In vitro assays demonstrated that synonymous codon changes in flanking regions can modulate MAP presentation.
• Cross-species and cell-type applicability: Applicable across multiple cell types and species.

Scientific Applications:

• Immunopeptidome prediction: Improve identification of peptides presented by MHC-I for mapping the immunopeptidome.
• Vaccine and immunotherapy target prioritization: Aid prioritization of MAPs relevant to virus-infected and neoplastic cells for CD8 T cell responses.
• Complementary predictor integration: Combine with MAP ligand scores and transcript expression levels to refine MAP repertoires.
• Study of translational regulation: Investigate the role of codon usage and arrangement in the biogenesis of MAPs.

Methodology:

Uses an artificial neural network trained on mRNA codon sequences flanking MCCs (excluding MCCs), compares original versus shuffled codon sequences that preserve amino acid usage, and integrates CAMAP scores with MAP ligand scores and transcript expression levels.