Probing T-cell response by sequence-based probabilistic modeling.
Probing T-cell response by sequence-based probabilistic modeling.
复制标题
用基于序列的概率模型探测T细胞应答。
DOI：
10.1371/journal.pcbi.1009297
复制
发表时间：
2021-09
影响因子：
4.3
通讯作者：
Cocco S
中科院分区：
文献类型：
--
作者：
Bravi B;Balachandran VP;Greenbaum BD;Walczak AM;Mora T;Monasson R;Cocco S
随着使用高通量下一代测序来量化人T细胞受体（TCR）库的多样性的能力的增加，使用TCR序列来推断抗原特异性的能力可以极大地帮助潜在的诊断和治疗。在这里，我们使用一种被称为受限玻尔兹曼机的机器学习方法来开发一种基于序列的推理方法来识别抗原特异性TCR。我们的方法结合概率模型的TCR序列与克隆丰度信息提取TCR序列基序中央抗原特异性反应。我们使用该模型来识别对个体肿瘤和传染病抗原有反应的患者个性化TCR基序，并准确区分特异性和非特异性反应。此外，模型的隐藏结构导致可解释的表示空间，其中TCR响应于相同的抗原簇，正确区分TCR对不同病毒表位的响应。该模型可用于识别条件特异性响应TCR。我们专注于TCR反应的候选新抗原和刺激TCR克隆扩增实验中选择的表位的例子。大量免疫细胞，如T细胞，越来越多地通过高通量测序获得。利用这些数据集来推断T细胞如何对抗原做出反应，可以帮助设计疫苗和过继性T细胞疗法。我们在这里提出了一种基于概率机器学习的方法来识别和表征响应T细胞。在学习之后，这种方法能够区分对不同抗原刺激有特异性反应的克隆。模型参数和T细胞序列的低维表示在分子水平上识别T细胞识别的序列基序。该方法说明了在体外刺激T细胞的癌症相关的新抗原，以及常见的感染性疾病的数据库数据。
With the increasing ability to use high-throughput next-generation sequencing to quantify the diversity of the human T cell receptor (TCR) repertoire, the ability to use TCR sequences to infer antigen-specificity could greatly aid potential diagnostics and therapeutics. Here, we use a machine-learning approach known as Restricted Boltzmann Machine to develop a sequence-based inference approach to identify antigen-specific TCRs. Our approach combines probabilistic models of TCR sequences with clone abundance information to extract TCR sequence motifs central to an antigen-specific response. We use this model to identify patient personalized TCR motifs that respond to individual tumor and infectious disease antigens, and to accurately discriminate specific from non-specific responses. Furthermore, the hidden structure of the model results in an interpretable representation space where TCRs responding to the same antigen cluster, correctly discriminating the response of TCR to different viral epitopes. The model can be used to identify condition specific responding TCRs. We focus on the examples of TCRs reactive to candidate neoantigens and selected epitopes in experiments of stimulated TCR clone expansion. Large repertoires of immune cells, such as T cells, are increasingly made available by high-throughput sequencing. Exploiting such datasets to infer how T cells respond to antigens could help design vaccines and adoptive T-cell therapies. We here propose an approach based on probabilistic machine learning to identify and characterize responding T cells. After learning, this approach is able to distinguish clones that specifically respond to different antigen stimulations. The model parameters and the low-dimensional representations of the T-cell sequences identify sequence motifs underlying T-cell recognition at the molecular level. The approach is illustrated on repertoire data describing in vitro stimulation of T cells by cancer-related neoantigens, as well as on data for common infectious diseases.
登录
查看更多内容
影响因子：
9.3
作者：
Bravi B;Tubiana J;Cocco S;Monasson R;Mora T;Walczak AM
通讯作者：
Walczak AM
影响因子：
7.3
作者：
Bjerregaard AM;Nielsen M;Jurtz V;Barra CM;Hadrup SR;Szallasi Z;Eklund AC
通讯作者：
Eklund AC
影响因子：
5.8
作者：
Sethna, Zachary;Elhanati, Yuval;Mora, Thierry
通讯作者：
Mora, Thierry
影响因子：
7.7
作者：
Davidsen, Kristian;Olson, Branden J.;Matsen, Frederick A.
通讯作者：
Matsen, Frederick A.
影响因子：
4.3
作者：
Puelma Touzel, Maximilian;Walczak, Aleksandra M.;Mora, Thierry
通讯作者：
Mora, Thierry