How to analyze TCR-repertoire from raw data of full length RNA-Seq SmartSeq2 ?

Question

[maxim.durand.etu]

Hi,

A platform will perform for our lab, in the next months, a full length RNA-Sequencing on T-cells and I wonder if it is possible to analyse TCR repertoires from the raw data (sorry for my ignorance in bioinformatics and genomics).

I saw that most of the TCR V beta analysis are based on a pre-amplification of the V beta sequences before sequencing analysis and really don't know if it could be done with non-amplified RNA-sequencing.

Thanks,
M.

Answer 1

[mikhail.shugay]

Dear Maxim,

You perform the analysis of RNA-Seq data with conventional RepSeq analysis software (have a look at MiTCR), but expect very low yield, as to fully characterize a T-cell clonotype one needs to obtain a complete CDR3 (V-J junction) sequence*. Given a sufficient sequencing depth you can get some insight on the repertoire composition.

I would recommend coupling RepSeq analysis with raw read mapping to V/J reference sequences using BWA. This will tell you if there is a significant difference between V-J segment usage and pairing in raw data compared to recovered TCRbeta clonotypes.

Also keep in mind recently published TCRklass software (paper) which is claimed to properly handle incomplete TCR sequences.

* none of recent tools allow assembly of CDR3 sequences, except for iSSAKE (paper) developed to handle very short reads

[maxim.durand.etu]

Hi Mikhail,

Thank you for your help, I think I will need to get closer to Geneticists / Bioinformaticians able to understand how to use analysis software you presented.

B.R.

Maxim

[mikhail.shugay]

Good luck then. But I think that RNA-seq data is not the best place to start getting familiar with Immune repertoire sequencing (Rep-Seq). I recommend you to get some publicly available RepSeq datasets (TCR ones) and polish your pipeline by comparing your RNA-Seq results to them as a "gold standard".