1
$\begingroup$

I have read a couple of publications related to the question being asked here, and have concluded that quantification of isoforms for scRNASeq experiments is typically coupled with long-read technologies, in order to overcome the limitations associated with short-read sequencing.

There is also a previous question asked on the forum which addresses this question (check: Detect transcript isoform abundance for a specific gene in scRNA-seq), but my question differs in the following way:

Rather than looking at the 'abundance' or quantity of different isoforms, I was wondering if I would be able to accurately identify the dominant isoform, in a cell-specific fashion of course. This reduced complexity may make the task solvable in situations where the quantification task is not solvable.

I was thinking this might be doable by generating contig sequences from reads from each of the cells, and then aligning it with the different isoforms and see which one matches the most.

Is this achievable ? Does anyone have an idea of how this can be done ?

Thanks.

$\endgroup$
2
  • $\begingroup$ You said your question differs from quantifying different isoforms (which includes discovering the most abundant isoform as a subset of the solution); what is that difference? $\endgroup$
    – gringer
    Commented May 21 at 4:48
  • $\begingroup$ The difference here I suppose, would be that I only wish to identify the most abundant isoform, rather than look at how other isoforms (including the most abundant) are being expressed. $\endgroup$
    – h3ab74
    Commented May 21 at 11:44

1 Answer 1

2
$\begingroup$

I don't know how you would define a dominant isoform, to me that is just the most abundant.

Anyway, what you are trying to achieve, I believe, is not possible with the 5' chemistry (or with the 3' for that matter). 10X gene expression libraries provide sequence information either for the 5' or the 3' the transcript and unless you are interested isoforms that differ in these regions, you will no be able find any evidence on different isoforms of a given gene. Quoting from 10X:

After amplifying the cDNA, molecules are randomly fragmented under conditions that favor 300-400 bp length fragments. Downstream of fragmentation, only transcripts containing both (1) a 10x Barcode AND (2) an Illumina Read 2 adaptor, which is ligated on to the cDNA after fragmentation, will be amplified during the Sample Index PCR. This results in final 10x libraries that either represent the 3' end of the transcript (as the 10x Barcode is adjacent to the polyA tail on the 3' end of the transcript) or the 5' end of the transcript (as the the 10x Barcode is adjacent to the TSO and the 5' end of the transcript).

$\endgroup$
1
  • $\begingroup$ Long-read approaches skip the fragmentation step, so that each read is a full representation of the transcript. $\endgroup$
    – gringer
    Commented May 18 at 20:02

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.