Filtering "merge.gtf" (from cuffmerge) by redundance

[Ramon Vidal]

I am trying to filter my merge.gtf file produced by ~100 samples by redundancy, I would like to remove the transcripts/isoforms that occur in only one sample. But I didnt find a way to track this information (where each transcript come from) easily!

Thanks a lot!

[Ramon Vidal]

Ok, I solved my problem just by using cuffcompare instead of cuffmerge.

[turnersd]

Here's another thread on this topic of whether to use cuffmerge or cuffcompare. Cole Trapnell and the cufflinks team is soon releasing a protocol paper that will hopefully address some of these issues.

You might think twice about using cuffcompare over cuffmerge.

[Ramon Vidal]

Thank you turnersd, actually I am using the second approach described on that thread "assemble each bam and cuffcompare them to get a combined.gtf".