TCR Sequencing (TCR-seq): Bulk vs Single-Cell Methods

Short answer. TCR sequencing (also written TCR-seq, or T cell receptor sequencing) reads the V(D)J-rearranged genes that encode the α and β chains of T cell receptors, so a sample’s T cell clones can be identified and counted. Bulk methods read very many receptors one chain at a time; single-cell methods keep the α/β pair and can add gene expression.

This page is for researchers and lab staff choosing or interpreting a TCR repertoire assay. After reading it you should be able to match a method and input type to your question and know what to record so the result can be reproduced. It is general method information, not clinical advice.

Definition and scope

A T cell receptor is built from an α chain and a β chain, each assembled from V and J gene segments (the β chain also uses a D segment) plus a constant region. The hypervariable CDR3 region sits at the centre of peptide-MHC recognition and is the usual identifier of a clone. The theoretical diversity is about 2×10¹⁹ αβ pairs, although any individual’s repertoire is far smaller (Pai and Satpathy, Nature Methods 2021).

TCR-seq therefore answers questions about the repertoire: which clonotypes are present, how abundant each is, and, for some methods, which α chain pairs with which β chain. It is distinct from single-cell RNA sequencing, which measures gene expression, although the two can be combined on the same cells.

How TCR sequencing works in practice

Bulk TCR-seq

Bulk approaches amplify receptor transcripts or genes from a pooled sample, using multiplex PCR or 5′ RACE, and sequence them. They capture very large numbers of receptors but resolve one chain at a time, “at the cost of resolving the pairing” of the α and β chains (Pai and Satpathy, 2021). They are scalable and cost-effective with simpler analysis, but do not preserve receptor chain pairing or cellular context (Seo and Choi, Genomics & Informatics 2025). Multiplex amplification is covered in multiplexed PCR.

Single-cell TCR-seq

Single-cell methods recover paired α and β chains, either by physically isolating cells (FACS into plates, droplets, or emulsion overlap-extension PCR) or by computationally pairing barcoded bulk data (pairSEQ). They can be combined with gene expression, surface protein or chromatin accessibility readouts. The trade-off is that they are more costly, more technically demanding and lower in throughput (Seo and Choi, 2025). Tools such as TraCeR, scTCRseq, VDJPuzzle and TRAPeS reconstruct TCRs from single-cell RNA-seq data (Pai and Satpathy, 2021); for the cost side, see single-cell sequencing cost.

Antigen specificity

Specificity can be mapped with fluorescent or DNA-barcoded peptide-MHC multimers (for example TetTCR-seq and MATE-seq). Rare antigen-specific cells can be as infrequent as about 1 in a million (Pai and Satpathy, 2021).

Choosing a method: a sourced decision table

Question you are asking	Approach	Why (source)
How many clones, and how large is each?	Bulk, genomic DNA input	gDNA is stable, includes productive and nonproductive rearrangements, and suits clone quantification, but carries no expression information (Seo and Choi, 2025).
Which clones are actively expressed?	Bulk, RNA input	RNA reflects the actively expressed repertoire but is less stable and is subject to extraction and reverse-transcription bias (Seo and Choi, 2025).
A more stable template than RNA?	Bulk, cDNA input	cDNA is more stable than RNA but retains the same transcriptional biases (Seo and Choi, 2025).
Which α chain pairs with which β chain?	Single-cell	Bulk data do not preserve chain pairing; single-cell methods do (Pai and Satpathy, 2021; Seo and Choi, 2025).
What are the clone’s phenotype and receptor together?	Single-cell, with expression or protein readout	Single-cell TCR-seq can be combined with transcriptome, surface protein or chromatin accessibility (Pai and Satpathy, 2021).

Analysis outputs

Reads are grouped into clonotypes, typically by CDR3, with an abundance for each. Common summaries are diversity metrics (Shannon entropy, Simpson’s index, richness), V(D)J gene usage, CDR3 length and clustering of similar receptors (Seo and Choi, 2025). UMI-based error correction improves accuracy (Pai and Satpathy, 2021). Published pipelines and tools include MiXCR and TRUST4 (pipelines), IMGT/HighV-QUEST and IgBLAST (annotation), VDJtools (post-processing), TCRdist and TCRMatch (clustering), and Immunarch and Alakazam (analysis packages). Raw reads are normally deposited in repositories such as the Sequence Read Archive.

What to record for reproducibility

The following checklist follows from the points above; each item changes how a result should be read.

• Input type: genomic DNA, RNA or cDNA, and the amplification approach (multiplex PCR or 5′ RACE).
• Chain or chains sequenced, and whether α/β pairing was recovered.
• Whether UMIs were used for error correction.
• Region covered: CDR3 only or full-length receptor.
• Annotation and clonotype-calling tool, with version.
• Diversity metric used, and the sequencing depth it was computed at.
• Reporting against the MiAIRR standard for sharing repertoire data (named in Pai and Satpathy, 2021).

Where it matters for labs

The method choice sets what data a project can produce and what it will cost in sample preparation, sequencing depth and analysis effort. Single-cell workflows are more costly and lower in throughput than bulk (Seo and Choi, 2025), so the decision is best made against the question in the table above before samples are collected, since input type cannot be changed afterwards. Sequencing itself is described in next-generation DNA sequencing.

Common misunderstandings

• Bulk TCR-seq can tell you which α goes with which β. It cannot; pairing requires single-cell or computational pairing methods.
• DNA and RNA clone frequencies mean the same thing. DNA counts rearranged genes including nonproductive ones; RNA reflects expressed repertoire and carries reverse-transcription bias.
• CDR3 alone fully describes the receptor. CDR3-only sequencing misses CDR1 and CDR2 context; full-length sequencing costs more and gives lower coverage per clonotype (Seo and Choi, 2025).
• More reads always mean more real diversity. Sequencing errors can create false diversity (Seo and Choi, 2025).

Frequently asked questions

What is T cell receptor sequencing?

It is the sequencing of rearranged TCR genes or transcripts to identify T cell clones and their abundance. TCR-seq and T cell receptor sequencing refer to the same family of methods.

What is the difference between bulk and single-cell TCR-seq?

Bulk reads many receptors from a pooled sample, one chain at a time, cheaply and at scale. Single-cell keeps α/β pairing and can add expression data, at higher cost and lower throughput.

Should I sequence DNA or RNA?

Use DNA to quantify clones; use RNA or cDNA to see the expressed repertoire, accepting extraction and reverse-transcription bias.

Sources

• Pai JA, Satpathy AT. High-throughput and single-cell T cell receptor sequencing technologies. Nature Methods 2021;18(8):881-892. PMC9345561 (read 2026-11-08).
• Seo K, Choi JK. Comprehensive analysis of TCR and BCR repertoires. Genomics & Informatics 2025. PMC11853700 (read 2026-11-08).