EUVIMEDEUVIMED entdecken
Open Access · CC BY-NC 4.0

Disrupting explicit encoding paradigms: property-interactive transformers decode T-cell receptor specificity beyond dataset biases

Luming Yang iD, Haoxian Liu, Alec Calanche, Sohret M Gokcek, Vishal Singh, Nicholas Sansoterra, Munir Akkaya, Billur Akkaya, Alper Yilmaz iD

Briefings in Bioinformatics · 2025

DOI: 10.1093/bib/bbaf609 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Briefings in Bioinformatics, 2025. Verfasst von Luming Yang und 8 weiteren Autor:innen. Beteiligte Einrichtung: Photogrammetric Computer Vision Lab., The Ohio State University, 2070 Neil Ave, Columbus, OH 43210, United States. Hauptthemen laut MEDLINE: Receptors, Antigen, T-Cell, Peptides. Die Arbeit nennt 80 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY-NC 4.0).

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

The human immune response relies on the unique ability of T-cell receptors (TCRs) to specifically bind to peptides, a process essential for immune surveillance and response. Although deep learning methods for prediction of TCR–peptide binding have proliferated, many encoder-based approaches learn dataset biases, greatly overestimating the model results, and ignoring the biochemical mechanisms and spatial properties affecting binding. Through our analysis, we found that interaction pairs generated by cross-mapping the amino acid properties between TCR and peptide implicitly simulate spatial structure, enabling machine learning models to capture information more effectively. Based on this insight, we developed T-cell receptor cross (TCRoss), a transformer-based model for large-scale learning. In addition, we observed that incorporating environmental information into the dataset not only mitigates learning biases but also improves performance. Experiments show that TCRoss consistently outperforms existing models in both observed contexts and de novo peptide scenarios. Wet-lab validation using T-cell activation assays confirmed the model’s predictions for nonbinding peptides and provided critical experimental evidence for model assessment. Biophysical validation confirms that high-attention residue pairs correspond to crystallographically observed binding interfaces.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Receptors, Antigen, T-CellPeptidesHumansProtein BindingMachine LearningComputational BiologyDeep LearningT-Lymphocytes

Stichwörter der Autor:innen

T-cell receptors (TCRs)cross-mappingpeptide bindingspatial structuretransformer

Substanzen

Receptors, Antigen, T-CellPeptides

Autor:innen und Institutionen

  • Photogrammetric Computer Vision Lab., The Ohio State University, 2070 Neil Ave, Columbus, OH 43210, United States.
  • Department of Computer Science and Engineering, Hong Kong University of Science and Technology, Kowloon, Hong Kong SAR, China.
  • The Ohio State University Wexner Medical Center, 460 W 12th Avenue, Columbus, OH 43210, United States.

ORCID-Profile: Luming Yang, Alper Yilmaz

Zitieren

APA

Yang, L., Liu, H., Calanche, A., Gokcek, S. M., Singh, V., Sansoterra, N., Akkaya, M., Akkaya, B., & Yilmaz, A. (2025). Disrupting explicit encoding paradigms: property-interactive transformers decode T-cell receptor specificity beyond dataset biases. Briefings in Bioinformatics. https://doi.org/10.1093/bib/bbaf609

Vancouver

Yang L, Liu H, Calanche A, Gokcek SM, Singh V, Sansoterra N, et al. Disrupting explicit encoding paradigms: property-interactive transformers decode T-cell receptor specificity beyond dataset biases. Briefings in Bioinformatics. 2025. doi:10.1093/bib/bbaf609

BibTeX
@article{Yang2025,
  author = {Yang, Luming and Liu, Haoxian and Calanche, Alec and Gokcek, Sohret M and Singh, Vishal and Sansoterra, Nicholas and Akkaya, Munir and Akkaya, Billur and Yilmaz, Alper},
  title = {Disrupting explicit encoding paradigms: property-interactive transformers decode T-cell receptor specificity beyond dataset biases},
  journal = {Briefings in Bioinformatics},
  year = {2025},
  doi = {10.1093/bib/bbaf609}
}
Kostenlos starten →