Decoding the interactions between T cell receptors and peptide-MHC

项目概要 T 细胞受体 (TCR) 对同源肽-主要组织相容性复合物 (pMHC) 的识别至关重要 适应性免疫识别。这种交互的某些特征是很好理解的，包括许多规则 控制肽与 MHC 的结合。然而，我们对三元 TCR:pMHC 复合物进行建模的能力仍然有限 出于三个主要原因：（1）数据可用性； (2) 约束力； (3)交叉反应性。在阐明规则时 TCR:pMHC 接口运作，这些努力旨在解决核心的基本问题 适应性免疫识别，具有重要的理论和实践意义，包括潜在的 具有选定特异性的新型受体的正向设计，最近涌入的 TCR 的“解码” 特定抗原靶标的测序数据，以及对交叉反应潜力的了解 剧目。之前，我们开发了新方法，为构建 预测 TCR 特异性各个方面的算法 (1)，包括我们称为 TCRdist 的算法 - 一个简单的算法 以及比较 TCR 序列的有效距离测量。 TCRdist 可用于对抗原特异性进行聚类 TCR 序列，可以合并到基于距离的分类器中，能够正确分配 以前未观察到的 TCR 对特征库具有强大的敏感性和特异性。综合起来， 这些实验的结果和 TCRdist 算法的普遍成功提供了令人信服的证据 对于这个提议的中心前提：给定足够数量的经过实验验证的表位- 特定的TCR序列，可以从其序列预测TCR的表位特异性； 此外，表位特异性 TCR 序列数据的生成，结合结构 知情的计算分析，为构建 TCR:pMHC 的预测模型提供了路线图 互动。虽然我们在这方面的调查中取得了重大进展，但剩下的最大障碍是 库内明显广泛的交叉反应性。为了充分阐明复杂的网络 TCR 和 pMHC 之间的相互作用，我们必须解决的问题是：不同的 TCR 看到了什么？ 相同的 pMHC 有共同点吗？相同 TCR 看到的不同 pMHC 有何作用？ 常见的？除了直接的生物学应用之外，这些研究的最终结果将是 协助开发用于 TCR:pMHC 相互作用建模的下一代分析工具， 从而实现这个基本接口的真正“解码器”的最终目标。

PROJECT SUMMARY T cell receptor (TCR) recognition of a cognate peptide-major histocompatibility complex (pMHC) is central to adaptive immune recognition. Certain features of this interaction are well-understood, including many of the rules governing peptide binding to MHC. However, our ability to model the ternary TCR:pMHC complex remains limited for three primary reasons: (1) Data availability; (2) Binding; and (3) Cross-reactivity. In elucidating the rules by which the TCR:pMHC interface operates, these efforts stand to address fundamental questions at the heart of adaptive immune recognition, with important theoretical and practical implications that include the potential for the forward design of novel receptors with selected specificities, the “decoding” of the recent influx of TCR sequencing data for specific antigenic targets, and an understanding of the cross-reactive potential of the repertoire. Previously, we developed novel approaches that provided training data for the construction of algorithms that predict various aspects of TCR specificity (1), including an algorithm we call TCRdist - a simple and effective distance measure to compare TCR sequences. TCRdist can be used to cluster antigen-specific TCR sequences and can be incorporated into a distance-based classifier capable of correctly assigning previously unobserved TCRs to characterized repertoires with robust sensitivity and specificity. Taken together, the results of these experiments and the general success of the TCRdist algorithm provide compelling evidence for the central premises of this proposal: Given a sufficient number of experimentally verified epitope- specific TCR sequences, the epitope specificity of a TCR can be predicted from its sequence; furthermore, the generation of epitope-specific TCR sequence data, in combination with structurally informed computational analysis, provides a roadmap for building a predictive model of the TCR:pMHC interaction. While we have made significant progress in this line of inquiry, the largest remaining hurdle is the apparent broad cross-reactivity within the repertoire. In order to fully elucidate the complex network of interactions among TCRs and pMHCs, the questions we must address then are: what do diverse TCRs that see the same pMHC have in common? And what do diverse pMHCs that are seen by the same TCRs have in common? The ultimate consequence of these studies, beyond their immediate biological applications, will be to assist in the development of the next generation of analytical tools for the modeling of TCR:pMHC interaction, leading to the ultimate goal of a true “decoder” for this essential interface.