This is a test version of Biostars. For the public version, visit https://www.biostars.org.
How to select isoform which shows the highest expression within a gene?

Dear Biostars Community,

I'm trying to correct for degraded samples using the 3' tag counting published method (https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0091851).

"The 3TC method involves changing the GTF annotation file, that is used when counting, in two steps; i) isoform filtering and ii) transcript length restriction. The isoform filtering step selects which isoform of each gene to retain as to prevent count contamination from overlapping isoforms. "

As the article was from 2014 they used Cufflinks to select the most expressed isoform within a gene. Currently I'm struggling in finding the current reliable method to obtain a similar result, as I was told that there are several more recent methods. I made a test with StringTie, But I'm sure is the best fit for my need.

Thank you in advance for any suggestion.

Best regards,

Youssef

stringtie cufflinks star rna-seq isoform

If anyone will need the full pipeline to get to the result, contact me in private

1 answer

I would use salmon to quantify reads against the transcriptome. That will give abundances and TPM per isoforms that you can conveniently filter.

Kallisto or RSEM can do that too. I think Kallisto was published after 2014.

Thank you very much ATpoint, it took me some times but I get the result with salmon :)

Log in to answer this question.