Is it possible to get more unique number of clonotypes per sample than t-cells?

[misrak]

We used the following library "Chromium Single Cell 5' Reagent Kits User Guide (v2 Chemistry Dual Index)" enriched with VDJ. I used the built-in preset

mixcr -Xmx60g analyze 10x-sc-xcr-vdj -s hsa {input.reads1} {input.reads2} {sample_name}
mixcr exportAirr {params.sam}.clna {params.sam}.airr.tsv

I used immunarch to estimate the number of unique clonotypes per sample. And, if I had 10,000 cells for a sample then I got 20,000 number of unique clonotypes. I feel this is an over-estimation. I think I am missing something critical here.

One more question, can I use mixcr to estimate the number of t-cells in a sample?

Thank you for the help.

[mizraelson]

It depends on how you define a clonotype. In any case, you can directly count the number of lines in the output file or the number of cells with two chains (if you define a clone as a pair of chains). There is no need to use Immunarch for this.
MiXCR can only work with TCR/BCR alignments, so you can count the number of cell barcodes that have TCRA/TCRB chains. However, if your question is about using gene expression for that, MiXCR does not support it yet.

[misrak]

Okay. Which output file should I use to count the number of t-cells and number of unique clonotypes? The command I used to analyse the samples is correct?

[mizraelson]

The command is correct. You can use the clones.tsv file. The number of lines is the number of clones. The number of unique CELL barcodes is the number of cells identified.