Hi everyone,

I am interested in getting your thoughts on an issue related to the comparison of event- and transcript-based approaches to analyzing alternative splicing. Im using event (splicing) based analysis and transcript based analysis with the descriptions from this post in mind.

Let's suppose you perform an event-based analysis and obtain PSI values for certain splicing events in your sample. The PSI value tells you what percent of isoforms contain the event in question. However, if your gene of interest has, for example, 3 annotated isoforms and 2 of them contain the event in question, there is no way to know about the relative levels of the 2 isoforms, right?

Does this then justify performing a transcript-based analysis alongside the event-based analysis? This would be done so you can look at the expression of the annotated transcripts and determine exactly which transcript containing the event in question is more highly (or lowly) expressed. Let's say, for the sake of the argument, that you would use salmon and take a quasi-mapping approach to get reads that map to the annotated transcripts (from a reference transcriptome).

Let me know if I can provide more information to make my question clearer. I am including a schematic to illustrate my points since it helped me wrap my head around the problem. If you could make some suggestions or provide any readings you think are useful, that would be greatly appreciated.

https://ibb.co/1sN69Pd - link to schematic

Cheers,

Fjodor