what does miRNA* mean?

Question

[l0o0]

I am reading a paper about miRNA and confused by a fig.

In "Novel miRNA for which miRNA* were found", what does miRNA* mean in the context? Does it mean reverse-complement sequence of miRNA?

I hope someone could give me some suggestions.

[Brian Bushnell]

The asterisk refers to the line at the bottom of the image, "*Known miRNAs refer to..."

[Michael]

That is misleading within this figure, they should have used a different character.

[ddiez]

The * points to the clarification at the end of the figure. I quote:

*Known miRNAs refer to miRNAs with <=2 mismatches to miRNAs deposited at the miRBase v20."

That is, from the "novel miRNAs" some are similar to "known miRNAs" (according to the criteria specified in *).

Answer 1

[Michael]

miRNA* refers to the antisense sequence to the mature miRNA. miRDeep classifies miRNA reads based on the location of reads in the pre-miRNA: miRNA (mature miRNA sequence), miRNA* (as sequence of miRNA), and loop (matching loop region of pre-miRNA hairpin).

[Asaf]

It's a bit deprecated these days and miRNAs are referred to as 5'-miRNA and 3'-miRNA since it was hard to determine which side is the active one.

[ddiez]

Oh, I see. Missed that detail (not familiar with miRDeep). Would you please provide a link where this is documented/explained? Thank you.

[Michael]

Check out the documentation or check the PDF generated for each Mirna. There they use the same notations

[l0o0]

Thank yourfor your answer. In my mind, novel miRNAs seem to be not in the miRbase but they have a miRNA-like structure. Is it necessary to check miRNA* for novel miRNA?

[Michael]

• Sure, novel miRNA are not in any database. However, if you don't provide mirDeep2 with information on known miRNA, it will summarize all detected miRNA as 'novel'.
• running mirDeep alone is not sufficient to classify short RNA into miRNA, and other small RNA. Running cmscan (earlier Rfam_scan) is a good way to get additional annotation information.
• The mirDeep algorithm does check for miRNA, miRNA*, and loop; and counts the numbers of reads mapping to either. I think that the expectation is that miRNA should get the highest number of reads, because of how the miRNA is integrated with RISC. If detecting miRNA de-novo, this helps to determine which part of the stem of the hairpin is used.