STAR parameter --outSAMstrandField

Pawan Chinari

unread,

Mar 25, 2020, 2:38:27 PM3/25/20

to rna-star

Dear Prof. Dobin,

I have a query regarding one of the parameters while using STAR for aligning RNA-seq data.

I have stranded RNA-seq data and I did the alignment using STAR v2.7.2a. Following alignment, I want to use the Leafcutter software for downstream processing.

In the leafcutter manual page (http://davidaknowles.github.io/leafcutter/articles/Usage.html), they suggest the use of “--outSAMstrandField intronMotif” in “Step 0. Alignment” of the manual.

From the STAR manual, it is suggested to use this parameter only for un-stranded RNA-seq data.

I did the alignment for a few samples with and without the parameter “--outSAMstrandField intronMotif” and the results differ slightly.

Since the data I have is stranded RNA-seq data, I wanted to ask your suggestion as what would be the best solution ?

Thank you for your help.

Best Regards,

Pawan

Alexander Dobin

unread,

Mar 25, 2020, 5:31:25 PM3/25/20

to rna-star

Hi Pawan,

I do not have experience with LeafCutter, but if they suggest using this parameter, I would go with their suggestion.

It adds the strand tag to spliced reads based on the splice motifs, so it does not require strandedness of the libraries.

Possibly, alternatively, you can add strand flag to every read based on the strandedness of your library - but I would inquire with the Authors of LeafCutter of this is a good idea.

The mapped differences with --outSAMstrandField intronMotif are small - basically it filters out alignments with non-canonical unannotated junctions.

Cheers

Alex

Pawan Chinari

unread,

Mar 26, 2020, 4:22:13 PM3/26/20

to rna-star

Dear Prof. Dobin,

Thank you for your response.

Does it make a major difference with and without this parameter ?

Why it is suggested not to use this parameter when the data is stranded RNA-seq ? I am asking because the data I have stranded RNA-seq.

Thanks,

Pawan

Alexander Dobin

unread,

Mar 30, 2020, 11:41:36 AM3/30/20

to rna-star

Hi Pawan,

Does it make a major difference with and without this parameter ?

The mapped differences with --outSAMstrandField intronMotif are small - basically, it filters out alignments with non-canonical unannotated junctions.

Why it is suggested not to use this parameter when the data is stranded RNA-seq ? I am asking because the data I have stranded RNA-seq.

Is this suggested for LeafCutter? Other software (e.g. Cufflinks) have their own options for stranded libraries, so this flag is not needed.