Quantification of abundance of one gene splicing isoform

Question

[aimanbarki]

Hello,

Sorry for the long post but i am lost and need suggestions I want to use publicly available human-RNA seq data to know which splicing isoform of certain gene s more abundant in that data and which one is less abundant. I don't know how to pull that information from RNA seq data. What I tried so far is below:

I downloaded some RNA seq data ( t-RNA >200bp) and use hisat2. Then open these bam file in IGV and make sashimi plot. but the issues with this approach are the following.

1. Gene is big and have long introns which make visual in IGV difficult.
2. Introns also show alot of aligned reads. ( So could it be because i am using t-RNA? As The alignment quality in the introns area is mostly 60)
3. Is this is the right approach to use sashimi plot to estimate the abundance of the splicing isoform ?
4. I look for other approaches and read about the MISO (looks abandon, no recent updates)or ggsashimi (just make plot but does not provide quantification). please guide me towards most recent and well developed source which i can use.

Thanks

Answer 1

[rpolicastro]

If you want transcript level quantification you should use Salmon. There are a few considerations required when quantifying at a transcript level that older methods of gene level quantification don't address, so you need to use software designed specifically for this purpose. Speaking generally in short read sequencing you won't have direct evidence of what isoform many reads were derived from due to many isoforms sharing exons. This requires you to build a model in order to infer (relatively speaking) what proportion of reads were likely derived from what isoforms.

[aimanbarki]

I am sorry I new to this. Can you please explain "This requires you to build a model in order to infer (relatively speaking) what proportion of reads were likely derived from what isoforms." How can i do this?

[rpolicastro]

Salmon does the work for you, you just need to run the software. As always I recommend reading the paper too to get an idea of what's going on under the hood.

[aimanbarki]

Hey I have run the Salmon followed by DRIMSeq and StageR using the following tutorial. I am only interested in one gene and it appears in the output files of stage R. but, padaj value for that gene is NA. Can please explain what is this means.