I tried to predict some 18S rRNA structure of ciliate following recent publications. Especially for the variable region 2 and 4 of 18S rRNA. However, I couldn't get the same with the published results. Depending on the length of input sequence, the results were highly variable from the mfold and RNAstructure.
Could you inform me of a more detailed guide or manual to predict the structure? For example, how to prepare the input seq. The input seq should have stem region in begining of 5' and ending of 3' as a pair?
Hi, predicting a reliable secondary structure is still an open task. The accuracy of the prediction decreases with the length of the sequence. Moreover, different programs use different strategy to fold the molecule so I am not surprise that you get different results using mfold or RNAstructure. You can try to use a local programming fold such as RNALfold and compare the results in your region of interest (region 2 or 4) with the published results. In these results you are referring to, do the authors use evolutionary information? Thi type of information will change the output of the program.
The input sequence does not require preprocessing,copy and paste the sequence is sufficient.
Your best bet is doing structural homology modeling. The secondary structure of rRNA is very well known and highly conserved troughout both kingdoms, so if you group sequences of know rRNA with your sequence, any one of these programs that do structural homology should be able to predict the structure with some reasonable reliability; particularly if you collect a diversity of sequences from various phyla.
I would say it is a waste of time to do these predictions by thermodynamics without any homology; in principle they _should_ predict correctly, but basically, there are fundamental things wrong with all of them: including my own approach (though for quite different reasons than other approaches).
At any rate, programs that try to fit structures to known structures (structural homology modeling) will do a far better job of providing you with what you need. You might try checking this site: http://its2.bioapps.biozentrum.uni-wuerzburg.de/ which is dedicated to rRNA. It seems a little hard to get used to, but your sequence may already be basically there. If you want to look of particular comparative sequence alignment programs, this is a _somewhat_ complete list.... http://en.wikipedia.org/wiki/List_of_RNA_structure_prediction_software#Comparative_methods
You might try looking at Gutell's database
http://www.rna.icmb.utexas.edu/
He is one of _the_ rRNA folk.
Hi Jae-Ho,
I am not an expert at predicting 2o structures myself, but I have a former postdoc and another research collaborator who are really good at this. I am sure they would be glad to help you over the rough spots. Please send me an email message through ResearchGate so I can put you in touch with them by email (both are in China).
You can use also the Vienna RNA package: http://www.tbi.univie.ac.at/RNA/
In this package you can use RNAalifold that calculate secondary structures for a set of aligned RNAs. Or RNALalifold that calculate locally stable secondary structures for a set of aligned RNAs.
Thanks all the people gave me the kind comments. Unfortunately, I'm still in problem. The 2nd structure I have attached is predicted from the sequence mentioned below. If someone know the way for prediction of 2nd structure, please help me.
TCGAATTTACATGGATAACCGTGGTAATTCTAGAGCTAATACATGCTGTTTAGCCCGACTTTTCGGAAGGGCTGTATTTATTAGATAACAAGCCAATATTCCTTCGGGTCTATTGTGATGATTCATAATAACTGATCGAATCGCATGGACTTTGTCCGCGATACATCA
- Badges
- Science topic
- Similar topics
- Biological Science
- Molecular Biology
- Biomolecules
- Nucleic Acids
- RNA