Identify the isoform of a transcript using RNA Seq data

Question

[aggregatibacter]

Hi,

my favorite transcript is supposed to consist of 12 exons (it was only described in germ cells, not in somatic tissues). Thus, I want to check what isoforms are actually expressed in my (somatic) tissue of interest.

I have paired-end RNA Seq data from a larger number of samples. I would like to use this dataset to shed light onto the isoform.

Thus far, I have aligned my reads using STAR, and counted exons using featureCounts. Seemingly, only a few exons are expressed, but quite consistent over all samples. There is no described isoform of the transcript that looks like what I have.

How would you suggest to go from here to obtain a consensus sequence? Do I need to assemble the whole transcriptome, or are there any shortcuts I can take, as I am only interested in a single transcript?

Many thanks!

Answer 1

[Devon Ryan]

I'd subset the alignments to contain only that regions and then run stringTie.

[aggregatibacter]

That sounds like a great idea, many thanks, Devon!

Answer 2

[Manvendra Singh]

Did you try running Cufflinks2 RABT assembly; I would suggest you to run in both mode; default and RABT and see if you can catch your desired isoform

Answer 3

[EagleEye]

These posts might give you some idea,

A: How to determine alternative splicing read counts

A: workflow/tool for alternative splicing