The number of peaks in microRNA-Seq trimmed data

Question

[wangdp123]

Hi there,

From the adaptor-trimmed single-end microRNA-Seq datasets, I observe that there are three peaks in terms of the read length.

1) peak 1 at 11bp;

2) peak 2 at 22bp;

3) peak 3 at 43bp.

The 22bp peak should indicate the enrichment of microRNA species.

But how about 11bp peak and 43bp peak? Is it normal to have three peaks for microRNA sequencing? Where do the peak1 and peak3 come from?

Is it important to select the read length, for example, only choose read length of 18bp-26bp for the downstream analysis such as quantification and differential expression analysis? If so, which tool can be used to select the read length?

If read length selection is done, it appears only a small proportion of the total reads from the experiment are from microRNA species. Is this generally accepted? What percentage would make sense?

Many thanks,

Tom

Answer 1

[jperezboza]

I think is common protocol to discard reads for small RNA sequencing below 15-16nt, most trimmers come with an option for that (trimmommatic and cutadapt for example).

Regarding your 3rd peak, that is most likely linked to piRNAs (~30nt). Where do you see more reads? In peak 1,2 or 3? If you are using cellular RNA this is normal, check out in my paper "Exploring the RNA landscape of endothelial exosomes" and in Fig 2A you can see how I also found 2 peaks in cellular RNA at the same size you have yours.