TCRsift¶
Prioritize T-cell receptor clonotypes from paired single-cell VDJ and gene expression data.
Overview¶
TCRsift combines observed clonal expansion, T-cell phenotype and expression state, public-database annotation, sequence publicness, and optional TIL matching. These are prioritization signals: antigen specificity still requires experimental validation.
TCRsift takes standard 10x Genomics Cell Ranger outputs as input:
- VDJ output from
cellranger vdj, containing TCR contigs and clonotypes - Matching GEX output from
cellranger count, when the workflow uses CD4/CD8 phenotyping or gene-expression signatures
Using those inputs, it can:
- prioritize expanded clonotypes with biology-aware filtering
- score T-cell phenotype and published expression signatures
- annotate known database matches and sequence publicness
- find culture-enriched TCRs in matched tumor samples
- assemble full-length TCR sequences for selected candidates
VDJ-only workflows are also supported when expression-based analyses are not needed. See Input requirements for the expected files and sample-sheet fields.
Quick example¶
Install TCRsift and create a minimal sample sheet pointing to the Cell Ranger output directories:
samples:
- sample: "Patient1_Culture"
vdj_dir: "/data/patient1/vdj"
gex_dir: "/data/patient1/gex"
Run the pipeline:
The result directory contains per-cell data, clonotype tables, plots, and a
record of the resolved configuration. Database annotation and sequence assembly
outputs appear when their optional inputs are provided. source defaults to
culture. TIL-only analyses use source: "til" and the
multi-sample TIL workflow, not this
culture-oriented run command.
See the Quick Start for a complete example or the Python API for step-by-step control.
Choose a workflow¶
| Starting data | Start here |
|---|---|
| Antigen-stimulated culture, optionally with matched TIL | Quick Start |
| Multiple TIL VDJ + GEX samples | Multi-sample TIL Prioritization |
| A single-cell atlas that needs QC, embedding, and cell typing | Single-Cell Atlas Path |
Installation¶
The quick example uses the core PyPI package. See Installation for optional dependencies and source installation.