Is there a way to run Cufflinks' cuffcompare in a strand-agnostic mode?
I would like to do this because I have some RNA-seq datasets derived from an unstranded run, that should be compared to a reference transcriptome. I think that it makes no reason to have some transcripts in the experimental dataset classified with transfrag code tag x and s as compared to the reference if the input dataset was not derived from stranded protocol.
From manual:
x : Exonic overlap with reference on the opposite strand
s : An intron of the transfrag overlaps a reference intron on the opposite strand (likely due to read mapping errors)
Conceptually, does it make sense?
If so, and if the run by cuffcompare cannot be changed to a strand-agnostic manner, how should I interprete those transcripts classified as x? I was thinking of merging them to o and/or c .
Again, from manual:
c : Contained
o : Generic exonic overlap with a reference transcript
