How can I best sequence T and B cells for the sequence of their receptors?

Question

[pasolol]

Hi guys,

I would like to profile the transcriptomic profile of T and B cells. I am most interested in the sequence of T and B cell receptors. I would like to use a plate-based protocol like SMARTSeq2 or MARSSeq2.

Do you have suggestions how I can do that?

Do I necessarily have to go for a full-length protocol or is it also possible to use a 3' or 5' protocol?

Do I need to use a bait-based enrichment to capture BCR or TCR sequences or are there bioinformatic tools that can get me the RNA sequence from a 'standard' NGS run where I amplify all sequences evenly?

Thanks a lot for your input!

Answer 1

[mizraelson]

Hi, it really depends on the task. First of all - yes, you can use SMARTSeq2 protocol. These protocols involve fragmentation so you usually get full-length at the end anyway, which might be a good thing especially for B-cells due to presence of hypermutations throughout the entire V-gene. These protocols don't really suggest BCR and TCR targeting and usually are used to create cDNA libraries that fully cover entire genes from polyA to UTR. That means that the yeild is gonna be low. If your main interest is in the TCR and or BCR repertoires I would suggest thinking of a protocol with enrichment. Yes, there are bioinformatic tools that work with total RNA-seq data. I am personally one of the developers of MiXCR software. We support SmartSeq2 protocol out of the box:

mixcr analyze smartseq2-vdj-full-length \
-s hsa \
161014_SN163_0729_AH3VW7BCXY_L1_{{CELL0ROW:a}}-{{CELL0COL:a}}_R1.fastq.gz \
161014_SN163_0729_AH3VW7BCXY_L1_{{CELL0ROW:a}}-{{CELL0COL:a}}_R2.fastq.gz \
result

https://docs.milaboratories.com/mixcr/reference/overview-built-in-presets/#smart-seq2

I can also help you with parameters to use if you end up using some other protocol. Please do check the link above - it leads to the list of protocol supported by MiXCT already.

Answer 2

[Matthias Zepper]

As far as the bioinformatic part is concerned, the nf-core pipelines are usually a good choice to analyze data according to established best practice standards - in your case AIRRflow would probably be the right choice. In their Slack community, there is usually a separate channel for every pipeline to reach out to the developers regarding questions that you can't answer by reading though the data specification.

For the web lab part, I unfortunately can't tell you anything. Maybe there is a good protocol available on protocols.io ?