A Public Benchmark for TCR-pMHC Prediction

Accurate prediction of T cell receptor (TCR) specificity for cognate peptides presented by the major histocompatibility complex (pMHC) is a key challenge in understanding adaptive immunity and enabling immune-based therapeutics and diagnostics. Successful models predicting TCR-pMHC specificity must be able to generalize to novel TCRs and novel pMHCs not observed during model training. Here, we provide a labeled dataset intended for general use as a test of TCR-pMHC model generalization to novel TCRs specific for novel pMHCs.Dataset file: "immrep2025_for_release.tsv"Detailed dataset description file: "immrep25_dataset_and_methods.pdf"

精准预测T细胞受体（T cell receptor, TCR）对主要组织相容性复合体（major histocompatibility complex, MHC）所呈递的同源肽的结合特异性，是解析适应性免疫机制、开发基于免疫的治疗与诊断手段的核心挑战。能够实现TCR-pMHC结合特异性预测的优质模型，需具备对训练阶段未观测到的新型T细胞受体与新型pMHC复合物的泛化能力。本研究提供一款适用于通用场景的带标注数据集，用于测试TCR-pMHC结合特异性模型对靶向新型pMHC复合物的新型T细胞受体的泛化性能。数据集文件为："immrep2025_for_release.tsv"，详细数据集说明文件为："immrep25_dataset_and_methods.pdf"