Criteria for defining MIRNA and other types of sRNA

[SergioRodLla]

Hello,

First of all thank you for developing ShortStack pipeline, I'm new to these world of analyzing small RNA-seq data and I found your tool very useful for the identification/annotation step. I would like to ask you about which is the current criteria for considering a locus to be a MIRNA according to the pipeline. Does this only apply to plant microRNA annotation as you discussed in your paper? I'm using mouse small RNA data. Why can the "Length" of the locus in the Results.txt file be bigger than ~24 nt and still be considered as a MIRNA? How does the length relate to the number of reads of each length reported for every locus? (I'm probably getting this idea wrong). Finally, would you recommend to use this tool for identifying other types of sRNA such as tsRNA, siRNA, mRNA, ...? My goal is to identify the proportion of each type that I have in my samples, and I wonder if ShortStack can be suitable if I provide other databases in the --knownRNAs parameter.

Thanks a lot in advance! Sergio

[MikeAxtell]

I would like to ask you about which is the current criteria for considering a locus to be a MIRNA according to the pipeline. Does this only apply to plant microRNA annotation as you discussed in your paper? I'm using mouse small RNA data.

• Yes the current ShortStack release uses the "plant rules" to distinguish MIRNA loci from other sRNA loci.

Why can the "Length" of the locus in the Results.txt file be bigger than ~24 nt and still be considered as a MIRNA? How does the length relate to the number of reads of each length reported for every locus? (I'm probably getting this idea wrong).

• "Length" in Results.txt is the length of the entire locus. For a MIRNA locus, this is the hairpin plus/minus about 20nts on either end.

Finally, would you recommend to use this tool for identifying other types of sRNA such as tsRNA, siRNA, mRNA, ...? My goal is to identify the proportion of each type that I have in my samples

• ShortStack will de novo find any locus with a good pile of aligned small RNA reads to your genome. The data in "Results.txt" should help you do the categorization (along with things like bedtools intersect to correlate to other annotations in your genome).

and I wonder if ShortStack can be suitable if I provide other databases in the --knownRNAs parameter.

• As of release 4.0.2, the --knownRNAs option has been renamed to --known_miRNAs to make its function more clear. This parameter is only meant to enhance MIRNA locus finding by using the alignment positions of user-determined mature miRNAs to nucleate searches for loci that meet MIRNA accumulation and secondary structure criteria. It is not suitable for other types of sRNA annotations. You can always just align other RNAs you have to your genome and use bedtools intersect to correlate your alignments with ShortStack's sRNA loci.

[SergioRodLla]

Thank you for the clarifications!