CrossTope

Structural Database for Cross-Reactivity Prediction in MHC:Peptide Complexes

WHY | The Challenge

Cellular immunity relies on T-cell receptors (TCR) recognizing peptides presented by Major Histocompatibility Complex (MHC) molecules. A critical—yet poorly understood—phenomenon is cross-reactivity: one TCR can recognize multiple different peptides bound to the same MHC allele. This mechanism drives heterologous immunity and molecular mimicry, but predicting which peptides trigger cross-reactive responses remains challenging.

The gap: Existing structural databases lack experimental validation, reproducible methods, and integrated immunogenicity data, limiting their application in vaccine design and immunotherapy.

WHAT | Our Solution

CrossTope is a highly curated repository of three-dimensional structures of peptide:MHC class I complexes (pMHC-I), combining crystallographic data with large-scale in silico modeling using our validated D1-EM-D2 approach.

Key differentiators:

  • Experimental validation required: Every epitope must demonstrate CTL response in vitro and/or in vivo
  • Manual curation: Each complex includes curated immunogenicity data and cross-reactivity relationships verified from literature
  • Continuously expanding: Growing collection of non-redundant complexes across human (HLA-A02:01, HLA-B27:05) and murine (H-2-Db, H-2-Kb) alleles
  • Integrated access: Direct links to IEDB, NCBI, and UniProt for comprehensive epitope information

Available data:

  • Structural coordinates (.pdb format)
  • Topological and charge distribution maps of TCR-interacting surfaces (−5/+5 and −10/+10 kT)
  • Immunological background with cross-reactivity evidence

HOW | Methodology & Applications

D1-EM-D2 Structural Modeling:

Our approach combines molecular docking (D1), energy minimization (EM), and refinement docking (D2) to reproduce crystallographic structures with high accuracy (mean RMSD 0.882 Å for Cα atoms). This method successfully identified molecular features explaining immunogenicity variation among HCV-derived epitope variants, revealing shared charge distribution patterns in immune-stimulating complexes.

Methodological references:

Practical applications:

  • Cross-reactivity prediction: Cluster structurally similar epitopes to identify potential cross-reactive targets
  • Viral escape analysis: Evaluate mutations at structural level to understand immune evasion
  • Immunotherapy optimization: Improve tumor antigen immunogenicity through rational design
  • Vaccine development: Identify conserved structural features across pathogen variants

Experimental validation: Our structure-based predictions identified cross-reactive HCV epitope variants with minimal sequence identity, later confirmed experimentally with lymphocytes from infected patients and vaccinated individuals—demonstrating the power of structural over sequence-based approaches.

Access & Integration

Team

Gustavo Fioravanti Vieira

Gustavo Fioravanti Vieira

Main Investigator

Marcelo A. S. Bragatte

Marcelo A. S. Bragatte

Database Maintainer & Researcher

frontendwizard

frontendwizard

Developer

Research Group

NBLI (Núcleo de Bioinformática do Laboratório de Imunogenética) is a research team focused on identifying immunological mechanisms triggered by viral infections and potential viral targets.

Financial Support

Bill & Melinda Gates FoundationCNPqCAPES

How to Cite

Primary reference:

Sinigaglia M, Antunes DA, Rigo MM, Chies JAB, Vieira GF. CrossTope: a curate repository of 3D structures of immunogenic peptide: MHC complexes. Database (Oxford). 2013;2013:bat002. doi: 10.1093/database/bat002

Read the full paper

BibTeX:

@article{sinigaglia2013crosstope,
  title={CrossTope: a curate repository of 3D structures of immunogenic peptide: MHC complexes},
  author={Sinigaglia, Marialva and Antunes, Dinler A and Rigo, Maurício M and Chies, José AB and Vieira, Gustavo F},
  journal={Database},
  volume={2013},
  pages={bat002},
  year={2013},
  publisher={Oxford University Press},
  doi={10.1093/database/bat002}
}

PubMed: PMID: 23396301
PMC: PMC3567486

Related Projects