Predicting binding probability of circRNA and miRNA

Question

[ResearchR]

Dear all, I am working on circular RNA and miRNA. I used the tool DCC to get possible circular RNA candidates from my STAR alignment. Since it is know that circular RNA expression can effect miRNA in various ways (et vice versa), I would like to compute the likelihood that one specific circRNA binds to a miRNA (or vice versa). Since some of my circRNAs are not annotated yet, I can just use the genomic coordinates and the according sequence. It would be too simplistic to just compare the sequences of circRNA and miRNA, without taking spatial structure and other binding proteins into account. I have done quite some literature research, but I could not find any approach (mathematical nor data base), which deals with such a problem.

Does someone of you guys have a similar problem or a suggestion?

Best wishes

Answer 1

[IP]

Hi!

I have work a little on circRNA too, and as far as I know, there is no resource for doing what you want. However, in order to resolve your question, there are some web servers that could be helpful to address your problem:

I would like to compute the likelihood that one specific circRNA binds to a miRNA (or vice versa). Since some of my circRNAs are not annotated yet, I can just use the genomic coordinates and the according sequence.

As far as I know there are no tools for dealing with this, what I usually do is to take the circRNA sequences and use them as if they were mRNA sequences in order to scan them in the miRNA- mRNA binding prediction web servers. If your circRNA are annotated, you can use some tools like circinteractome to search for both RBPs and miRNA binding to your circRNAs.

It would be too simplistic to just compare the sequences of circRNA and miRNA, without taking spatial structure and other binding proteins into account.

The RNA.fold web server has an option were you can tell the software to assume that the RNA molecule is circular. This could allow you to have a computational overview of the posible folding structure of your molecule.

Bonus: Recent studies using ribosome footprinting have suggest that circRNA could code for proteins, it may be worthy for you to look for coding potential in your circRNA, see this nature research highlight were they explain this last experiments.