So I am performing a biomarker (piRNA based) study in a human solid cancer. Based on the initial analysis, I have shortlisted a few piRNAs - but I am stuck with two issues:
1. A few piRNAs have highly similar sequences - differing only by 2-3 terminal bases. Although the mechanisms of piRNA activity have not been clearly elucidated - I understand that similar to miRNAs, piRNAs also depend on a few core bases within their entire structure to act upon their targets. Considering this, it becomes a dilemma to choose between highly similar piRNA sequences. How do I proceed here? Should I be working with all of them, or just the parent sequence?
2. Another issue that has been bothering me is that a large majority of these piR sequences map to multiple chromosomal locations - making it a tad bit difficult to bioinformatically determine the potential targets of the piRNAs. Any leads here on how to proceed?
Thank you.
Hi Deepak Iyer ,
I am facing a similar problem, by trying to account for small RNA Seq reads and checking for piRNAs. It's the first time, working with these and the sequences from piRBase http://www.regulatoryrna.org/database/piRNA/download.html seem to be especially redundant. Clustering them with CD-HIT http://weizhongli-lab.org/cd-hit/ reduces the list down to ~80% (still clustering as I write) from their published 8,438,264 piRNA sequences. The latter number seems absurd to me (that's more twice than all possible 11mers, and thus not very specific for a CLASS of RNA). Their annotation is downright unhelpful, since single piRNAs are annotated up to 92,879 times throughout the genome. (please someone set me straight if I am misinterpreting any of these)
pirnadb.org has a set of 27,723 piRNAs (non-compatible IDs with piRBase) and is reduced to 22,028 piRNAs by CD-HIT. Annotation of these again seems to be meaningless with > 800k (again, please enlighten me otherwise).
The major downside of both these databases is my lack of functional connection of these piRNAs. piRBase gives some potential targets (not in human) and a total of 15 associations with breast cancer from 3 studies (which is a good start).
I'm new to piRNAs and in need of more resources myself. Thus I'd be eager to hear from others doing piRNA analysis. Most studies seem to treat piRNAs as everything that can not be accounted for by other ncRNAs in small RNA-Seq.
- Badges
- Science topic
- Similar topics
- Biological Science
- Biochemistry
- Biomolecules
- Nucleic Acids
- RNA
- Small RNA
- microRNA