PromPDD

PromPDD predicts and designs promiscuous peptides that bind human leukocyte antigen (HLA) class I molecules to identify epitopes with broad population coverage.

Key Features:

• ARB matrix-based prediction models: Uses the Average Relative Binding (ARB) matrix method to generate prediction models with a consistent, conservative threshold for peptide binding across well-studied HLA class I alleles.
• Residue-level binding deciphering: Deciphers the contribution of individual peptide residues at each position to promiscuous binding, revealing molecular determinants of broad HLA recognition.
• Design guidelines for promiscuous peptides: Provides ARB-based guidelines to design peptides capable of binding multiple HLA molecules for vaccine development and immunotherapy applications.
• Pan-specific matrices for limited-data alleles: Employs a derived modular method to generate pan-specific matrices enabling prediction for over 2,641 HLA molecules encoded by HLA-A and HLA-B genes, including alleles with limited experimental data.
• Supertypes and cross-allele prediction: Identifies peptides capable of binding multiple HLA alleles across or within supertypes.

Scientific Applications:

• Vaccine development: Identifies promiscuous peptides to design vaccines with wide population coverage by targeting peptides presented by multiple HLA class I alleles.
• Immunotherapy and personalized medicine: Supports selection and design of peptides for immunotherapy and personalized medicine by enabling broad HLA recognition.
• Peptide–HLA interaction research: Enables investigation of molecular determinants of peptide binding and promiscuity across HLA class I alleles.

Methodology:

Prediction models are generated using the Average Relative Binding (ARB) matrix method, and pan-specific matrices are derived via a modular method to extend predictions to alleles with limited experimental data.