UpTCR: a unified progressive knowledge transfer foundation model for robust T-cell receptor-antigen binding recognition
7 affiliations
- 1School of Artificial Intelligence and Computer Science, Jiangnan University, Wuxi, 214122 China
- 2Tsinghua Shenzhen International Graduate School, Tsinghua University, Shenzhen, 518071 China
- 3Cancer Centre and Institute of Translational Medicine, Faculty of Health Sciences, University of Macau, Macau, 999078 China
- 4AI for Life Sciences Lab, Tencent, Shenzhen, 518000 China
- 5Department of general surgery, Peking Union Medical College Hospital, Beijing, China
- 6Institute of Biomedical Engineering and Instrumentation, Hangzhou Dianzi University, Hangzhou, 310018 China
- 7The PRC Ministry of Education Engineering Research Center of Intelligent Technology for Healthcare, Wuxi, Jiangsu 214122 China
Abstract
T-cell receptor (TCR) recognition of antigenic peptides presented by human leukocyte antigen (HLA) molecules underpins adaptive immunity and T cell-based immunotherapy. However, the scarcity of complete interaction data and TCR cross-reactivity challenge robust prediction. Here, we present UpTCR, a unified progressive knowledge-transfer foundation model that learns from incomplete data to predict TCR-antigen-HLA binding. UpTCR progressively transfers knowledge from dimeric and trimeric interactions to tetrameric complexes and uses soft contrastive learning to mitigate false negatives. It outperforms existing methods in predicting TCR binding specificity and antigen-HLA binding affinity, particularly for neoantigens, and reveals pairwise residue-level interactions across the tetramer. UpTCR also transfers effectively to breast cancer cohorts with limited data. Prospective validation against melanoma antigen variants identifies eight immunogenic peptides that elicit T cell responses and one variant associated with immune escape. These findings establish UpTCR as a generalizable and interpretable tool for studying antigen recognition and advancing TCR-based immunotherapies.
Related
- Structural basis of phospholipid sensing nuclear receptor LRH-1 recognition by carboxylic acid bioisosteres reveals an alternative ligand interaction site
- A murine MHC-Ib molecule, H2-Q9, drives peptide-centric T-cell receptor recognition of a viral antigen
- Receptor-Specific Molecular Recognition of Maresin 1 and Docosahexaenoic Acid by the Nuclear Receptors PPARα, PPARγ, and RORα
- A Progressive 99mTc-MDP Uptake in Transthyretin Cardiac Amyloidosis
- The Progressive Cardiac Uptake of 99mTc-MDP as a Clue to Transthyretin Amyloidosis
BibTeX
@article{lv2026uptcr,
title = {{UpTCR: a unified progressive knowledge transfer foundation model for robust T-cell receptor-antigen binding recognition}},
author = {Tianxu Lv and Yang Xiao and Li Chen and Bing He and Maiyi Zhong and Zihan Feng and Zheyu Hu and Fei Ye and Jiashu Han and Shouzhi Chen and Zhenchao Tang and Jiale Zhou and Dawei Huang and Xiaoqing Lian and Jiansong Fan and Yixuan Huang and Chenyi Lei and Dandan Meng and Yuan Liu and Lihua Li and Pengjiang Qian and Jianhua Yao and Kai Miao and Xiao Liu and Xiang Pan},
journal = {Nature Communications},
year = {2026},
month = sep,
doi = {10.1038/s41467-026-78075-x},
url = {https://doi.org/10.1038/s41467-026-78075-x},
note = {CC BY 4.0},
}
RIS
TY - JOUR
TI - UpTCR: a unified progressive knowledge transfer foundation model for robust T-cell receptor-antigen binding recognition
AU - Tianxu Lv
AU - Yang Xiao
AU - Li Chen
AU - Bing He
AU - Maiyi Zhong
AU - Zihan Feng
AU - Zheyu Hu
AU - Fei Ye
AU - Jiashu Han
AU - Shouzhi Chen
AU - Zhenchao Tang
AU - Jiale Zhou
AU - Dawei Huang
AU - Xiaoqing Lian
AU - Jiansong Fan
AU - Yixuan Huang
AU - Chenyi Lei
AU - Dandan Meng
AU - Yuan Liu
AU - Lihua Li
AU - Pengjiang Qian
AU - Jianhua Yao
AU - Kai Miao
AU - Xiao Liu
AU - Xiang Pan
PY - 2026
DA - 2026/09/24
JO - Nature Communications
DO - 10.1038/s41467-026-78075-x
UR - https://doi.org/10.1038/s41467-026-78075-x
AB - T-cell receptor (TCR) recognition of antigenic peptides presented by human leukocyte antigen (HLA) molecules underpins adaptive immunity and T cell-based immunotherapy. However, the scarcity of complete interaction data and TCR cross-reactivity challenge robust prediction. Here, we present UpTCR, a unified progressive knowledge-transfer foundation model that learns from incomplete data to predict TCR-antigen-HLA binding. UpTCR progressively transfers knowledge from dimeric and trimeric interactions to tetrameric complexes and uses soft contrastive learning to mitigate false negatives. It outperforms existing methods in predicting TCR binding specificity and antigen-HLA binding affinity, particularly for neoantigens, and reveals pairwise residue-level interactions across the tetramer. UpTCR also transfers effectively to breast cancer cohorts with limited data. Prospective validation against melanoma antigen variants identifies eight immunogenic peptides that elicit T cell responses and one variant associated with immune escape. These findings establish UpTCR as a generalizable and interpretable tool for studying antigen recognition and advancing TCR-based immunotherapies.
SN - 2041-1723
ER -