RTCR: a pipeline for complete and accurate recovery of T cell repertoires from high throughput sequencing data.
RTCR: a pipeline for complete and accurate recovery of T cell repertoires from high throughput sequencing data.
复制标题
DOI：
10.1093/bioinformatics/btw339
复制
发表时间：
2016-10-15
期刊：
影响因子：
--
通讯作者：
de Boer RJ
中科院分区：
文献类型：
--
作者：
Gerritsen B;Pandit A;Andeweg AC;de Boer RJ
动机：高重复序列测序（HTS）使研究人员能够探测人类T细胞受体（TCR）库，其中包括许多罕见的序列。区分真正但罕见的TCR序列和由聚合酶链反应（PCR）和测序错误产生的变体仍然是一个艰巨的挑战。处理错误的常规方法是去除低质量读数和/或稀有TCR序列。这样的过滤丢弃了大量真实的并且通常罕见的TCR序列。然而，稀有TCR序列的准确鉴定和定量对于库多样性估计是必不可少的。结果如下：我们设计了一个称为Recover TCR（RTCR）的管道，即使在低覆盖率下，也可以从HTS数据（包括条形码数据）中准确地恢复TCR序列，包括罕见的TCR序列。RTCR采用数据驱动的统计模型，以自适应的方式纠正PCR和测序错误。使用模拟，我们证明了RTCR可以很容易地适应不同类型的测序仪的错误配置文件，并表现出一贯的高召回率和高精度，即使在低覆盖率，其他管道表现不佳。使用公布的真实的数据，我们表明，RTCR准确地解决测序错误，优于所有其他管道。可用性和实现：RTCR管道是用Python（v2.7）和C实现的，可以在http://uubram.github.io/RTCR/along上免费获得，并提供文档和典型用法的示例。联系人：B. uu.nl
Motivation: High Throughput Sequencing (HTS) has enabled researchers to probe the human T cell receptor (TCR) repertoire, which consists of many rare sequences. Distinguishing between true but rare TCR sequences and variants generated by polymerase chain reaction (PCR) and sequencing errors remains a formidable challenge. The conventional approach to handle errors is to remove low quality reads, and/or rare TCR sequences. Such filtering discards a large number of true and often rare TCR sequences. However, accurate identification and quantification of rare TCR sequences is essential for repertoire diversity estimation. Results: We devised a pipeline, called Recover TCR (RTCR), that accurately recovers TCR sequences, including rare TCR sequences, from HTS data (including barcoded data) even at low coverage. RTCR employs a data-driven statistical model to rectify PCR and sequencing errors in an adaptive manner. Using simulations, we demonstrate that RTCR can easily adapt to the error profiles of different types of sequencers and exhibits consistently high recall and high precision even at low coverages where other pipelines perform poorly. Using published real data, we show that RTCR accurately resolves sequencing errors and outperforms all other pipelines. Availability and Implementation: The RTCR pipeline is implemented in Python (v2.7) and C and is freely available at http://uubram.github.io/RTCR/along with documentation and examples of typical usage. Contact: b.gerritsen@uu.nl
登录
查看更多内容
影响因子：
56.9
作者：
Arstila, TP;Casrouge, A;Kourilsky, P
通讯作者：
Kourilsky, P
影响因子：
48
作者：
Langmead, Ben;Salzberg, Steven L.
通讯作者：
Salzberg, Steven L.
DOI：
10.1073/pnas.1001705107
发表时间：
2010-03-23
影响因子：
11.1
作者：
Mora, Thierry;Walczak, Aleksandra M.;Callan, Curtis G., Jr.
通讯作者：
Callan, Curtis G., Jr.
影响因子：
4.4
作者：
Klarenbeek, Paul L.;Tak, Paul P.;de Vries, Niek
通讯作者：
de Vries, Niek
影响因子：
48
作者：
Shugay, Mikhail;Britanova, Olga V.;Chudakov, Dmitriy M.
通讯作者：
Chudakov, Dmitriy M.