outSAMstrandField intronMotif with stranded data

[praful aggarwal]

Hi,

I accidentally used the '--outSAMstrandField intronMotif' parameter with stranded RNA-Seq data and ran cufflinks downstream without the '--library-type' option (Sorry I completely flipped what I read in the manual :D) and cufflinks ran fine and gave me an output. I am not sure if I completely comprehend how these two parameters work together (I am reading more about it), but do you think I can still trust the cufflinks output or should I re-run the whole thing again without the STAR parameter and by telling cufflinks the library type? Also, can we even use the "intronMotif" option for stranded data or is it wrong to do that? If we use it then how would it affect the strand information?

Thanks for any help.

Praful

[Alexander Dobin]

Hi Praful,

without --library-type, Cufflinks will only be able to determine strand for spliced reads, which I think will significantly reduce the quality of the assembly.

I would definitely recommend re-running Cufflinks with the --library-type. The main purpose of --outSAMstrandField intronMotif command is to add XS strand tags to spliced reads.

It also remove unannotated non-canonical junctions for which the strand cannot be determined, but this is a small effect - so I would not re-run STAR.

Cheers

Alex

[praful aggarwal]

Hi Alex,

Thank you for that reply. After the original post, as I was reading more about these I figured that Cufflinks will need to be re-done. Just to add to your point, I actually re-ran STAR w/o the --outSAMstrandField and compared the HTSeq count based differential gene expression results (for about 15 samples) using the w/ and w/o --outSAMstrandField intronMotif output and the results look "identical" (some genes have a few reads off) because I have always HTSeq that my data is stranded (default -s parameter) and I don't think that HTSeq really cares about the "XS" value. So not all is lost :)

Thanks again for all your help. 

Praful