Evolution of the tuberculin skin test reveals generalisable Mtb-reactive T cell metaclones

This dataset collection contains all de novo T cell receptor sequence (TCR) data, in processed format, used for analysis in our TST TCR manuscript (https://doi.org/10.1038/s41467-026-68678-9). A metadata file (<code>metadata.csv</code>) is included that provides annotations for each filename of individual samples. The data were generated with two different methods, and comprise several data subsets:<b>Dataset 1:</b> Bulk TCR sequencing data generated with the UCL Chain lab protocol. The raw data files for these samples will be available at the NCBI Sequence Read Archive, accession number PRJNA1208718, upon peer-reviewed publication of the manuscript. The processed data files, available here, were generated using Decombinator v4 (https://github.com/innate2adaptive/Decombinator).Alpha and beta chain sequences from whole blood of n=20 adults with latent tuberculosis (TB) infectionAlpha and beta chain sequences from skin punch biopsies taken 2-3 days after tuberculin skin test (TST) injection in n=17 adults with latent TB infectionAlpha and beta chain sequences from skin punch biopsies taken 7 days after TST injection in n=165 adults with latent TB infectionAlpha and beta chain sequences from cultured PBMC of n=12 adults with latent TB infection, stimulated for 6 days with one of 10 µg/mL purified protein derivative of Mtb, 100 µg/mL tetanus toxoid, or control buffer, yielding between 1-5 replicates per participant per stimulus.Individual sample data files are provided in <code>.tsv.gz</code> format. In addition, for convenience, combined data files of full or down-sampled single-chain repertoires are provided, which allow full reproducibility of the analyses presented in the manuscript, using analysis code on https://github.com/carolinturner/tst_tcr. Full repertoires = <code>combined_alpha.csv.gz</code> and <code>combined_beta.csv.gz</code>; Repertoires down-sampled to 16,000 TCRs each = <code>combined_subsampled_alpha.csv.gz</code> and <code>combined_subsampled_beta.csv.gz</code>; Repertoires down-sampled to between 5,000 and 10,000 TCRs for metaclone discovery = <code>combined_subsampled_5000_10000_beta.csv.gz</code>.<b>Dataset 2:</b> Bulk TCR sequencing data generated with Adaptive Biotechnologies' ImmunoSEQ assay. The raw data files for these samples are not made available by the third party provider. The processed data files, available here, were downloaded from the ImmunoSEQ website.Beta chain sequences from lung tissue resections of n=13 adults with TB disease. Up to 3 samples per patient, scored for disease severity by physician.Beta chain sequences from control lung tissue resections of n=3 adults with lung cancer without TB disease. Up to 2 samples per patient, including tumour and healthy margin samples.Beta chain sequences from blood of n=11 adults with TB disease and n=3 adults with lung cancer without TB disease. Up to 2 replicate samples per patient.Beta chain sequences from CD4 T cells, flow-sorted from lung or blood of n=5 adults with TB disease, and additionally separated by expression of tissue residency marker CD69.Individual sample files are provided in <code>.tsv</code> format. For convenience, a combined data file is provided for beta chain repertoires (<code>ImmunoSeq_combined_beta.csv.gz</code>).<br>

本数据集合集收录了用于本团队TST TCR研究论文（https://doi.org/10.1038/s41467-026-68678-9）分析的所有全新（de novo）T细胞受体（T cell receptor, TCR）序列数据，均为处理后格式。附带元数据文件（<code>metadata.csv</code>），可为每个单独样本的文件名提供注释信息。本数据集采用两种不同方法生成，包含多个数据子集：

**数据集1：** 采用UCL Chain实验室方案生成的批量TCR测序数据。本批次样本的原始数据文件将在论文经同行评议发表后，公开于NCBI序列读取档案库（NCBI Sequence Read Archive），登录号为PRJNA1208718。本平台提供的处理后数据文件，系通过Decombinator v4工具（https://github.com/innate2adaptive/Decombinator）生成。具体样本包括：
1. 来自20名潜伏性结核（latent tuberculosis, TB）感染成人全血的TCR α链与β链序列；
2. 来自17名潜伏性TB感染成人在结核菌素皮肤试验（tuberculin skin test, TST）注射后2~3天采集的皮肤活检组织的TCR α、β链序列；
3. 来自165名潜伏性TB感染成人在TST注射后7天采集的皮肤活检组织的TCR α、β链序列；
4. 来自12名潜伏性TB感染成人的培养外周血单个核细胞（peripheral blood mononuclear cell, PBMC）的TCR α、β链序列：这些PBMC经10 µg/mL纯化结核分枝杆菌蛋白衍生物（purified protein derivative of Mtb）、100 µg/mL破伤风类毒素，或对照缓冲液中的一种刺激培养6天，每名受试者每种刺激条件可获得1~5个生物学重复样本。

单个样本的数据文件以<code>.tsv.gz</code>格式提供。此外，为便于使用，本平台还提供完整或经降采样处理的单链免疫组库（repertoire）合并数据文件，可结合https://github.com/carolinturner/tst_tcr 上的分析代码，完全复现论文中呈现的分析结果。完整免疫组库文件包括<code>combined_alpha.csv.gz</code>与<code>combined_beta.csv.gz</code>；经降采样至每条样本含16,000条TCR序列的免疫组库文件为<code>combined_subsampled_alpha.csv.gz</code>与<code>combined_subsampled_beta.csv.gz</code>；用于元克隆发现、经降采样至每条样本含5,000~10,000条TCR序列的免疫组库文件为<code>combined_subsampled_5000_10000_beta.csv.gz</code>。

**数据集2：** 采用Adaptive Biotechnologies公司的ImmunoSEQ检测方法（ImmunoSEQ assay）生成的批量TCR测序数据。本批次样本的原始数据文件无法由第三方供应商公开提供。本平台提供的处理后数据文件，系从ImmunoSEQ官方网站下载获取。具体样本包括：
1. 来自13名活动性TB成人患者肺组织切除标本的TCR β链序列：每名患者最多可提供3份样本，由医师根据疾病严重程度进行评分；
2. 来自3名无TB感染的肺癌成人患者肺组织切除对照标本的TCR β链序列：每名患者最多可提供2份样本，包括肿瘤组织与健康边缘组织样本；
3. 来自11名活动性TB成人患者与3名无TB感染的肺癌成人患者血液的TCR β链序列：每名患者最多可提供2份重复样本；
4. 来自5名活动性TB成人患者肺组织或血液中经流式分选（flow-sorted）的CD4阳性T细胞的TCR β链序列：额外根据组织驻留标志物CD69的表达水平进行了分选分组。

单个样本的数据文件以<code>.tsv</code>格式提供。为便于使用，本平台还提供了β链免疫组库的合并数据文件（<code>ImmunoSeq_combined_beta.csv.gz</code>）。