Isoform level RNA-seq analyses

Question

[BPors]

Hi,

I am considering to filter my transcripts based on expression, in that case, what are the options to choose the highly expressed isoform of each gene in a count matrix ( for example, salmon or FeatureCounts)?

I found out that trinity has an option of isoform level filtering ( via its filter_low_expr_transcripts.pl ), but I was wondering, if there are other options you know of & trust in your analyses?

Thank you!

[ATpoint]

What is the analysis goal? Why do you want to filter?

[BPors]

I would like to keep the isoforms which are always being expressed regardless of the condition, so that I can use them in other analyses, for ex. checking how translation is effected etc.