The mFold input has a limit of 2400nt and while searching for common secondary strucutres around miRNA binding site of mRNA we have come across the issue of mRNAs with extremely long 3' UTRs, up to 4000nt long. This blocks our chosen strategy of assesing the entire UTR of each mRNA for secondary structures and modifying the amount of the RNA sequence has significant influence on the structures produced (for example 2000nt vs 2400nt of a UTR). The desciption of the methodology and rationale behind how much RNA is inputted mfold and why is very sparse in the literature.
Is there a resource for the defining basic rulesets for the selection of how much RNa to input into a system
The prediction of RNA secondary structures becomes less reliable with increasing sequence length; personally I would avoid predictions beyond 500 nt. In addition, for mRNAs, protein binding and ribosomes might hinder formation of thermodynamically optimal structures of the RNA.
Locally optimal secondary structures are a different thing: RNA structure formation is very fast, and the RNA folds (again) immediately after protein release or passage of a ribosome. Prediction of locally optimal structures is possible, for example, by RNALfold, RNAplfold, and RNALalifold from the Vienna package:
https://www.tbi.univie.ac.at/RNA/
You should install the mentioned programs on a local computer; to my knowledge no webservice is available for these.
To my knowledge, there is no general ruleset. As laid out by Gerhard Steger, you get more reliable predictions for shorter sequences. Unless you have good reason to fold the complete RNA sequence (e.g. you are interested in particular long-range interactions), I would go for not more than 200nt windows with RNALfold of the ViennaRNA package. BTW, the latest version of RNALfold has an option to backtrack a global MFE structure from locally optimal fragments via the '-b' option on the command line (no webservice available).
Thanks Gerhard Steger and Michael T. Wolfinger, very helpful advice I'll also link this paper attached which seems cover this issue to does a good job summarising and testing best practices for this exact problem.
Article Global or local? Predicting secondary structure and accessib...
- Badges
- Science topic
- Similar topics
- Biological Science
- Molecular Biology
- Biomolecules
- Nucleic Acids
- RNA