Dna Secondary Structure Prediction Of A Large Dataset

Question

[Timtico]

Hi,

I'm interested in prediction of hairpin structures from a large set of DNA sequences (fasta), which I extracted using my own perl scripts.

I'm looking for a tool that creates predictions of secondary structure (e.g. hairpin loops), and gives me some basic information such as free energy.

output should preferably in dot.bracket FASTA fileformat, so I can easily visualize the structures with known tools.

My weapons of choice would be either Perl or R. (I'm struggling with Taverna)

thanks in advance!

[Pierre Lindenbaum]

DNA ? not RNA ?

[Timtico]

yeah indeed DNA

[Cjt]

DNA secondary structure should be an ideal double strand, there isn't much to predict?! I assume you are looking for potential transcripts (haipin sounds like miRNA) by scanning genomic sequence.

[Aaronquinlan]

think single-stranded DNA.

[Timtico]

No i'm looking for specific hairpins surrounding a consensus sequence, indeed: think single stranded.

[Timtico]

indeed, think single stranded.

Answer 1

[Pasta]

I would use RNAfold from the Vienna RNA package which works for DNA too : http://rna.tbi.univie.ac.at/. You can either call the program from Perl or make a Perl version of the available source code.

As far as I can remember, you just supply a FASTA file and you will get dG and other thermodynamic values for each of your sequences.

[Timtico]

Thanks, I took a look at the code but i'm not very proficient in C. I found Unafold however, It's Perl based. I will try to make it work but running into some troubles.

Answer 2

[Asaf]

Try using RNAplfold from the same package, it predicts stable local structures that you can obtain from the dot-plot or use the -u option to get the probability for a region being unpaired. You can run the program and then use perl to digest the output

Answer 3

[Sequencegeek]

Hi There,

If you are interested in more detailed information about the hairpin, there is a script on this page that will parse the RNAFold output. It may be a little buggy but it has worked for me in the past. RNA fold Topic

Answer 4

[Timtico]

I ended up writing a wrapper/parser in Perl for the commandline executed RNAstructure package (University of Rochester) althought it's abit of a hack it works really well for my purposes.

The script takes a multiple sequence FASTA file and processes each sequence with fold.exe of the package. The results are gathered in dot.bracket FASTA file. From this, postscript images are created, depicting the predicted structure(s). Cutoff can be manually set to obtain only the structures with the highest likelyhood. All postscript files are gathered and output is combined into a single PDF-file using Ghostscript.