Researchers have harnessed the power of artificial intelligence to tackle one of the most complex challenges in immunology: predicting how T cells recognize and respond to specific peptide antigens. Using AlphaFold 3 (AF3), a AI/ML model, designed for protein structure prediction, the team demonstrated a novel approach to model T cell receptor–peptide/major histocompatibility complex (TCR-pMHC) interactions with growing accuracy.
T cells play a dual role in human health, acting as defenders by eliminating tumors and infected cells while sometimes contributing to disease by targeting the body’s own tissues. At the heart of this balance lies TCR-pMHC recognition, a critical process that determines whether T cells mount a protective response or trigger harmful autoimmunity. Until now, predictive models of TCR specificity have remained limited in accuracy and scope.
“Inspired by recent advances in AI-based structural biology, we sought to evaluate whether AlphaFold could be adapted to predict how T cells recognize epitopes,” said Dr. Chongming Jiang, Principal Investigator of the study. “Our findings indicate that AlphaFold can distinguish valid epitopes from invalid ones, moving us closer to reliable, high-throughput prediction of T cell responses.”
The research team reports that AlphaFold’s computational modeling enables in silico identification of immunogenic epitopes that could serve as vaccine targets. Beyond prevention, the ability to design higher-affinity and more specific T cells has the potential to enhance the safety and efficacy of T cell-based therapies for cancer, infectious diseases, and autoimmune conditions.
“An accurate prediction model of TCR-pMHC interactions could transform the development of immunotherapies and vaccines,” added Dr. Xiling Shen, Chief Scientific Officer at the Terasaki Institute. “This represents a crucial step toward precision medicine approaches that harness the immune system to combat disease.”
While the researchers acknowledge that further refinement and validation are required before widespread clinical application, the results highlight the promise of deep learning–based structural modeling as a pathway for the generalizable prediction of TCR-pMHC interactions.
This breakthrough underscores the potential of AI-driven approaches to accelerate drug discovery and immunotherapy design, paving the way for more effective and safer treatments.
Reference: Chao C chi, Chiu Y, Yeung L, Yee C, Jiang C, Shen X. AI/ML-empowered approaches for predicting T cell-mediated immunity and beyond. Front Immunol. 2025. doi: 10.3389/fimmu.2025.1651533
This article has been republished from the following materials. Note: material may have been edited for length and content. For further information, please contact the cited source. Our press release publishing policy can be accessed here.