Which reads are degraded mRNA?

[Joanne Harding]

Hi all,
I have a Illumina small RNA-seq dataset. I want to know if it is possible to distinguish between a genuine small RNA and a degraded messenger RNA fragment?

The dataset has already been compared to all known small RNA databases. I thought that the percentage of reads which align to exons of known genes might give an idea. Any other ideas/thoughts much appreciated!

[Lee Sam]

Hi Joanne. How about trying to align to both exons and splice junctions using short seeds and fairly tight thresholds (20mer, 1 mismatch)? You might end up with some reads that align as both small RNAs and degraded, but at least it'll give you an idea?

[steven]

Hi all,
I have a Illumina small RNA-seq dataset. I want to know if it is possible to distinguish between a genuine small RNA and a degraded messenger RNA fragment?

The dataset has already been compared to all known small RNA databases. I thought that the percentage of reads which align to exons of known genes might give an idea. Any other ideas/thoughts much appreciated!

Well, that is a whole debate..

If you assume that all small RNAs that are generated by cleavage of a longer mRNA are degradation products, then i guess that quite a few (if not all) miRNAs, snoRNAs or piRNAs for instance might technically be considered as "degraded mRNA fragment"..
In fact, thousands of sRNAs are made from mRNA cleavage, can be re-capped and polyAdenylated (PMID: 19169241).

In my humble opinion, the point is not just about the generation process ("degradation" vs. "processing"), but about the function of these molecules. Yet another debate: if you believe that only the known small RNAs can be functional and not the novel ones, then it is fair to consider the difference with the current databases as a good evaluation criteria -as novel sRNAs are still being reported these days, i doubt that this is a reasonable assumption. Otherwise, unless you adopt the simplistic "weak signal is noise" philosophy, i fear there is no easy way to address your question..