[joro]

Which versions of TopHat and Cufflinks are you using?

[ZhigangLi]

I used cufflink 2.0.2 and Tophat 2.0.6

[ravipatel4]

I did lot of research on cufflinks strand assignment, but could not get satisfactory information on it. It is really frustrating.

It makes sense that cufflinks assigns strand based on intron-exon junctions in case of multi-exonic transcripts. But, how it decides the strand for some of the single-exonic transcripts and puts "." for the rest? Could anyone find any answer on this?

I am surprised that the authors of cufflinks don't respond to this thread!

Please share any information you have on this.

Thank you for your attention.

[joro]

Hi ravipatel4,

On the Cufflinks website it says:
"Questions about Cufflinks should be sent to [email protected]"

Perhaps you could try that email address?

http://cufflinks.cbcb.umd.edu/

[ravipatel4]

Hi joro,

Thank you for your reply.
I already wrote to them. Hope to hear from them soon. I don't have much hopes, though. They might be getting many emails and they are more interested in emails related to bugs. Thus, at the first place, they might not be interested in respoding to my email asking for just a clarification.

It is strange that some papers already used cufflinks to assemble non-strand-specific RNA-Seq data, but they dint clarify anything on how they did proceed with the gene models without strand information!

Thank you.

[mendl]

We had an issue in this general area and I thought I might post what I learned. I had a novel single exon transcript called by Cufflinks but it consistently called the strand opposite of where the reads were mapping.(We were doing strand-specific sequencing.) We eventually figured out that Cufflinks uses the Tophat XS tag to call the strand of this feature and that the library type had been mis-set in the Tophat runs. e.g. Look for examples in your Tophat sam files where your read maps to the positive strand but the tag is "XS:A:-" or vice versa. Setting the library_type in Cufflinks did not fix this issue and eventually we corrected the mapping with a script. NOTE: transcripts in our .gtf annotation file were assigned the correct strand. We observed this issue for novel single exon transcripts.