Using your package having whole PBMC

[beginner984]

Hello all

I have only scRNA-seq 10x PBMCs of cancer free and controls.

My boss has asked me

T cell analysis – differences between samples (scRNAseq) Can we identify a/b or g/d pairs for individual T cells

Is there anyway to do TCR analysis just by just having scRNA-seq data from using your package or I need TCR sequencing data itself?

Thank you so much for any link, idea, tutorial, paper, etc

[ncborcherding]

Hey beginner984,

Thanks for reaching out.

As of right now, scRepertoire really only supports the paired scRNA/TCR sequencing approach. Based on my experience - just relying on RNA gets a little confusing, as single-cells RNA can show cells with multiple TCR loci - but it is do-able. Let me see if I can dig up a paper I saw, where they looked at the TCRBV genes by cluster and were able to get some degree of specificity by cluster.

Depending on your sequencing approach - you may be able to recover the TCR information using TRUST4. This approach works with 5' sequencing though, so it might not work for you.

I am going to close this "issue" because it is not a problem related to the code, however, please feel free to reach out to me anytime as you get started and I am more than willing to help.

Nick

[beginner984]

Hello

And thank you so much in advance

In our folders, I have found TCR/BCR sequencing for 3 out of 8 samples. Actually, by MiXCR we have extracted TCR and BCR CDR3 repertoires from our bulk RNA-Seq data (paired end fastq files as input). This has been done months ago before we run multiome 3' (scRNA-seq+scATAC-seq) q 10x PBMCs. I have attached header of TCR data for one sample here

cloneId cloneCount  cloneFraction   targetSequences targetQualities allVHitsWithScore   allDHitsWithScore   allJHitsWithScore   allCHitsWithScore   allVAlignments  allDAlignments  allJAlignments  allCAlignments  nSeqFR1 minQualFR1  nSeqCDR1    minQualCDR1 nSeqFR2 minQualFR2  nSeqCDR2    minQualCDR2 nSeqFR3 minQualFR3  nSeqCDR3    minQualCDR3 nSeqFR4 minQualFR4  aaSeqFR1    aaSeqCDR1   aaSeqFR2    aaSeqCDR2   aaSeqFR3    aaSeqCDR3   aaSeqFR4    refPoints
1   16.0    0.045845272206303724    TGTGCCAGTACCTCGACAGGGGATATGGTCACTGAAGCTTTCTTT   EEEEEEENNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNEEEEEEE   TRBV19*00(197.1)    TRBD1*00(40)    TRBJ1-1*00(62.8)    TRBC1*00(50.5)  427|437|464|0|10||50.0  14|22|36|14|22||40.0    24|40|68|29|45||80.0                                                TGTGCCAGTACCTCGACAGGGGATATGGTCACTGAAGCTTTCTTT   36                              CASTSTGDMVTEAFF     :::::::::0:-7:10:14:-2:-2:22:29:-4:45:::
3   14.0    0.04011461318051576 TGTGCCAGCAGCTTTTGGGGGGCGCCCCAGGATGAGCAGTTCTTC   EEEEENNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNEEEEEE   TRBV11-2*00(140)    TRBD1*00(55)    TRBJ2-1*00(45.7)    TRBC2*00(108.9) 430|444|467|0|14||70.0  18|29|36|17|28||55.0    28|42|70|31|45||70.0                                                TGTGCCAGCAGCTTTTGGGGGGCGCCCCAGGATGAGCAGTTCTTC   36                              CASSFWGAPQDEQFF     :::::::::0:-3:14:17:-6:5:28:31:-8:45:::
4   12.0    0.034383954154727794    TGTGCCAGTAGACACGAAAGTGGGACACAAAACAATGAGCAGTTCTTC    EEEEEENNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNEEEEEEE    TRBV19*00(195.4)    TRBD1*00(30)    TRBJ2-1*00(53.3)    TRBC1*00(35),TRBC2*00(35)   427|438|464|0|11||55.0  12|18|36|21|27||30.0    25|42|70|31|48||85.0    ;                                           TGTGCCAGTAGACACGAAAGTGGGACACAAAACAATGAGCAGTTCTTC    36                              CASRHESGTQNNEQFF        :::::::::0:-6:11:21:0:-6:27:31:-5:48:::
5   12.0    0.034383954154727794    TGTGCCAGCAGTGGTACTCAGACAGTGCGCACCACTGAAGCTTTCTTT    EEEEEEEEEENNNNNNNNNNNNNNNNNNNNNNNNNNNNNEEEEEEEEE    TRBV2*00(91.7)  TR

Now can I merge this TCR/BCR with my scRNA-seq using your software? Or please you know any other way in combining TCR/BCR comes from bulk RNA-seq to scrna-seq from the same sample?

[ncborcherding]

Hey thanks for the follow up!

Interesting that you recovered contigs using MixCR while using 3' scRNA-seq - the 5' chemistry is what is used for the VDJ sequencing, but I guess it is possible that a small fraction of reads would contain some VDJ component.

Unfortunately, MiXCR recovers the contigs without regards to cell barcode, so there is no way to attach it to the single-cell data. So MiXCR is essientially turning your single-cell into bulk-RNA in terms of contig recovery.

Nick

[beginner984]

Thank you so much

Let's say we have initially had bulk RNA-seq paired end fastq files (so we have had bulk RNA-seq rather than scRNA-seq)

We used these files as input for MixCR software to extract TCR/BCR CDR3 regions

For the same samples we have scRNA-seq so I was seeking for a way to relate clonotype information to cell types which likely does need cell barcodes

You think bulk RNA-seq deconvolution can help?

[ncborcherding]

You could try bulk-RNA deconvolution, but I think the simpler step would be to look at your single-cell data clusters for expression of the VDJ genes you recovered in the output of MiXCR. There is a large feature overlap (beyond TCR related genes) in T cells with shared clonotypes, so I bet you can get at least some insight into the "RNA phenotype" associated with the clones.