TCR Sequencing

T cell receptor sequencing reads the rearranged receptor genes of a sample and turns them into a census of T cells. Because the receptor sequence is fixed for the life of the cell, that census is also a lineage record: the same clonotype can be followed across timepoints, tissues and treatments, which is the main thing TCR data can do that BCR data cannot. This section works through the method from both ends, the wet lab that shapes what you can measure, and the analyses that turn annotated sequences into biology.

Every number, table and figure on these pages comes from a real run of the code shown on the page, inside one pinned container, on committed example data shaped exactly like the AIRR tables that nf-core/airrflow produces. The companion repository holds the runners, the fixtures and a small real excerpt of VDJdb.

The path

Page	What it covers
What is TCR sequencing	receptor structure, V(D)J recombination, thymic selection and clonotypes
Library Prep & Sequencing	which chain to sequence, amplicon PCR versus 5’RACE, UMIs, bulk versus single-cell
Repertoire Analysis	V(D)J assignment, CDR3 properties, clonal tracking, diversity and specificity prediction
Affinity Prediction Tools	how GLIPH2, TCRdist3, NetTCR and friends work, and where they fail
Running nf-core/airrflow	the pipeline for TCR data, plus immunarch and scirpy worked examples

The data

The worked examples run on three committed fixtures. Two AIRR-format TCR-beta tables stand in for airrflow output at two timepoints of one toy subject, and three of their expanded clonotypes are real, high-confidence VDJdb records, so the antigen-matching example finds something true. A synthetic 10x Genomics contig table feeds the single-cell scirpy example. The fixtures are deterministic, so every page here rebuilds bit for bit. When a page quotes a number or shows a figure, the container run beside the code produced it.