EUVIMEDOpdag EUVIMED
Open access · CC BY-NC 4.0

IMGT/RobustpMHC: robust training for class-I MHC peptide binding prediction

Anjana Kushwaha, Patrice Duroux, Véronique Giudicelli, Konstantin Todorov, Sofia Kossida

Briefings in Bioinformatics · 2024

DOI: 10.1093/bib/bbae552 ↗

At a glance

Journal article in Briefings in Bioinformatics, 2024. Written by Anjana Kushwaha and 4 co-authors. Institution: IMGT®, The International ImMunoGeneTics Information System®, Montpellier, France. Major MEDLINE topics: Histocompatibility Antigens Class I, Peptides, Computational Biology, Protein Binding. Cited 5 times so far (Crossref); 64 references. The full text is freely available (open access, CC BY-NC 4.0). Funded by: Centre National de la Recherche Scientifique.

AbstractSubjectsBibliografiske oplysningerCiteRelated studies

Abstract (original language)

The accurate prediction of peptide-major histocompatibility complex (MHC) class I binding probabilities is a critical endeavor in immunoinformatics, with broad implications for vaccine development and immunotherapies. While recent deep neural network based approaches have showcased promise in peptide-MHC (pMHC) prediction, they have two shortcomings: (i) they rely on hand-crafted pseudo-sequence extraction, (ii) they do not generalize well to different datasets, which limits the practicality of these approaches. While existing methods rely on a 34 amino acid pseudo-sequence, our findings uncover the involvement of 147 positions in direct interactions between MHC and peptide. We further show that neural architectures can learn the intricacies of pMHC binding using even full sequences. To this end, we present PerceiverpMHC that is able to learn accurate representations on full-sequences by leveraging efficient transformer based architectures. Additionally, we propose IMGT/RobustpMHC that harnesses the potential of unlabeled data in improving the robustness of pMHC binding predictions through a self-supervised learning strategy. We extensively evaluate RobustpMHC on eight different datasets and showcase an overall improvement of over 6% in binding prediction accuracy compared to state-of-the-art approaches. We compile CrystalIMGT, a crystallography-verified dataset presenting a challenge to existing approaches due to significantly different pMHC distributions. Finally, to mitigate this distribution gap, we further develop a transfer learning pipeline.

Subjects

MeSH terms (MEDLINE)

Histocompatibility Antigens Class IPeptidesComputational BiologyProtein BindingHumansNeural Networks, Computer

Author keywords

MHCdeep learningpeptiderobustself-supervisedtransformer

Substances

Histocompatibility Antigens Class IPeptides

Authors and institutions

  • IMGT®, The International ImMunoGeneTics Information System®, Montpellier, France.
  • Institute of Human Genetics (IGH), Montpellier, France.
  • University of Montpellier (UM), Montpellier, France.
  • National Center for Scientific Research (CNRS), France.
  • LIRMM, Laboratoire d'Informatique, de Robotique et de Microélectronique de Montpellier, Montpellier, France.
  • Institut Universitaire de France (IUF), Paris, France.

Funding

  • Centre National de la Recherche Scientifique

Cite

APA

Kushwaha, A., Duroux, P., Giudicelli, V., Todorov, K., & Kossida, S. (2024). IMGT/RobustpMHC: robust training for class-I MHC peptide binding prediction. Briefings in Bioinformatics. https://doi.org/10.1093/bib/bbae552

Vancouver

Kushwaha A, Duroux P, Giudicelli V, Todorov K, Kossida S. IMGT/RobustpMHC: robust training for class-I MHC peptide binding prediction. Briefings in Bioinformatics. 2024. doi:10.1093/bib/bbae552

BibTeX
@article{Kushwaha2024,
  author = {Kushwaha, Anjana and Duroux, Patrice and Giudicelli, Véronique and Todorov, Konstantin and Kossida, Sofia},
  title = {IMGT/RobustpMHC: robust training for class-I MHC peptide binding prediction},
  journal = {Briefings in Bioinformatics},
  year = {2024},
  doi = {10.1093/bib/bbae552}
}
Start for free →