Beyond Better Models: Barriers and a Path Forward in Generalisable TCR Antigen Specificity Prediction
- 1The University of Oxford
Abstract
T cells are at the centre of adaptive immunity, protecting the organism against infection and cancer. T cell receptors (TCRs), expressed on the surface of T cells, enable the recognition of target cells through interactions with peptides presented by major histocompatibility complex (pMHC) molecules. Reliable, large-scale mapping of TCRs to their cognate pMHCs remains a holy grail of modern systems immunology, with far-reaching implications for cancer immunotherapy, vaccine development, and the understanding and treatment of autoimmune diseases. While exhaustive experimental mapping remains intractable due to the enormous combinatorial sequence space of TCRs and pMHCs, computational approaches offer a promising alternative. Consequently, we have witnessed increasingly sophisticated computational models of TCR antigen specificity, from small neural networks to deep generative models and large protein language models. However, several independent benchmarks have shown that these models, despite their size and sophistication, largely fail to generalise to TCRs or pMHCs not present in the data available during model development. Here, we investigate possible reasons for this and suggest a path forward to generalisable TCR:pMHC prediction. First, our analysis of the largest public repositories of TCR:pMHC data reveals significant biases towards a small number of antigens and MHC molecules, sparse sampling of TCR sequence space per pMHC, as well as inconsistencies and inaccuracies in data collation. Second, we show the importance of carefully setting up model evaluation to measure discrimination based on antigen specificity instead of confounding factors present in the evaluation data. Finally, we highlight the need for and propose how best to match the evaluation metric with the specific downstream application of the predictions. Our investigation suggests that we, as a community, need more representative and better curated data, and a robust set of metrics and benchmarks to guide future model development towards generalisable TCR:pMHC specificity prediction.
Related
- General prediction of T cell receptor antigen specificity from sequence using AlphaFold 3
- Antigen specificity and inflammatory amplification in autoimmune hepatitis
- TCR origin and specificity to environmental or self-antigens on distinct antigen-presenting cells determine peripheral Treg cell differentiation
- Data coherence over data volume drives generalisable genome-based prediction of microbial carbon utilisation
- Limitations of general protein language models and public TCRpMHC data in training specificity prediction models
BibTeX
@misc{koohy2026beyond,
title = {{Beyond Better Models: Barriers and a Path Forward in Generalisable TCR Antigen Specificity Prediction}},
author = {H. Koohy and I. Ilievski and A. Antanaviciute and G. Ogg},
howpublished = {bioRxiv},
year = {2026},
month = oct,
doi = {10.64898/2026.10.12.758456},
url = {https://www.biorxiv.org/content/10.64898/2026.10.12.758456v1},
note = {CC BY-NC-ND},
}
RIS
TY - UNPB
TI - Beyond Better Models: Barriers and a Path Forward in Generalisable TCR Antigen Specificity Prediction
AU - H. Koohy
AU - I. Ilievski
AU - A. Antanaviciute
AU - G. Ogg
PY - 2026
DA - 2026/10/12
PB - bioRxiv
DO - 10.64898/2026.10.12.758456
UR - https://www.biorxiv.org/content/10.64898/2026.10.12.758456v1
AB - T cells are at the centre of adaptive immunity, protecting the organism against infection and cancer. T cell receptors (TCRs), expressed on the surface of T cells, enable the recognition of target cells through interactions with peptides presented by major histocompatibility complex (pMHC) molecules. Reliable, large-scale mapping of TCRs to their cognate pMHCs remains a holy grail of modern systems immunology, with far-reaching implications for cancer immunotherapy, vaccine development, and the understanding and treatment of autoimmune diseases. While exhaustive experimental mapping remains intractable due to the enormous combinatorial sequence space of TCRs and pMHCs, computational approaches offer a promising alternative. Consequently, we have witnessed increasingly sophisticated computational models of TCR antigen specificity, from small neural networks to deep generative models and large protein language models. However, several independent benchmarks have shown that these models, despite their size and sophistication, largely fail to generalise to TCRs or pMHCs not present in the data available during model development. Here, we investigate possible reasons for this and suggest a path forward to generalisable TCR:pMHC prediction. First, our analysis of the largest public repositories of TCR:pMHC data reveals significant biases towards a small number of antigens and MHC molecules, sparse sampling of TCR sequence space per pMHC, as well as inconsistencies and inaccuracies in data collation. Second, we show the importance of carefully setting up model evaluation to measure discrimination based on antigen specificity instead of confounding factors present in the evaluation data. Finally, we highlight the need for and propose how best to match the evaluation metric with the specific downstream application of the predictions. Our investigation suggests that we, as a community, need more representative and better curated data, and a robust set of metrics and benchmarks to guide future model development towards generalisable TCR:pMHC specificity prediction.
ER -