Deep-ComAIR: A Framework for Predicting TCR-pMHC Binding through Complex Structural Analysis
Keywords: AI for science, adaptive immunity, TCR-pMHC binding, multimodal integration
Abstract: The binding process between T cell receptor (TCR) and the peptide-major histocompatibility complex (pMHC) is a fundamental mechanism in adaptive immunity. Current research on binding prediction primarily emphasizes the sequence and structural features of critical regions within these molecules, often neglecting the intricate structural changes that occur at the binding process, which can lead to biased representations. To address this gap, we propose a novel framework, titled “Deep-ComAIR,” which effectively models the binding process by focusing on the complex structure of TCR-pMHC rather than individual components. This model enhances prediction accuracy by integrating features from three modalities: sequence, structural, and gene. Our approach achieves state-of-the-art results evidenced by an area under the receiver operating characteristic curve (AUROC) of 0.983 in binding reactivity prediction and a Pearson correlation coefficient of 0.833 in binding affinity prediction. These results highlight the framework's potential to deepen our understanding of TCR-pMHC interactions at the structural level and facilitate advancements in immunotherapy and vaccine design.
Primary Area: applications to physical sciences (physics, chemistry, biology, etc.)
Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics.
Submission Guidelines: I certify that this submission complies with the submission instructions as described on https://iclr.cc/Conferences/2025/AuthorGuide.
Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors’ identity.
No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.
Submission Number: 7044
Loading