miRNA analysis with bowtie, miRanalyzer or other

[hmortens]

I have just tested miRanalyzer and am now realizing that it may not use a very recent version of miRbase. The database construction is not trivial for this program however, and the bowtie indexes for the mature and hairpin screens are provided with the miRbase program. I am guessing that this is because the miRbase sequences are modified to align with bowtie.
I have tried making my own indexes and running bowtie with the miRbase fasta files, without success.
Does anyone have any experience with a good program to analyze miRNA, that reports known mature and hairpin structures from miRbase and can be updates regularly for use in a pipeline?
thanks,
holly

[Kennels]

Try the UEA sRNA workbench: http://srna-workbench.cmp.uea.ac.uk/

It has many tools, and can be run as GUI or command line.

[visserm]

Hi Holly
Have you added 25 G's after your mature miRNA sequences before you indexed them with Bowtie?
Marike

[hmortens]

thanks kennels and visserm. I actually got an update for miRbase 19 from the miRanalyzer folks, so it seems they are pretty on top of it.
Marike, regarding the bowtie indexing...Can you describe how you would do the alignment with miRbase fasta and mymiRNAdata.fastq? I basically constructed a bowtie index from the mature miRbase fasta and then ran bowtie with this index and my fastq files, which resulted in a bunch of junk. Do you have a script to add on the string of G's? and would you do this for the miRbase fasta before indexing?
many thanks.
holly

[visserm]

Hi Holly

I used miRanalyzer for analyzing known miRNA. As you know, miRanlyzer makes use of the bowtie indexed mature and hairpin files (the fasta files they got from mirBase). You can also index them (the mirBase fasta files) yourself and add them to the bowtie folder under miranalyzerDB. As I understand it, because the mature sequences are to short for bowtie analysis you have to add 25G's before indexing. I haven't used bowtie by itself to analyze known miRNAs, only via miRanalyzer.

Unfortunately I do not have a script to add the G's. I'm not to clued-up with bioinformatics.

Marike
PS mirBase20 is released. I also still need to index the new mirBase20 or get it from the miRanalyzer guys. They have a way of indexing all the miRNA from the different species separately

[empyrean]

to add 25 bp sequence, you can use the following command

awk 'NR % 2 {print} !(NR % 2) {print $0"GGGGGGGGGGGGGGGGGGGGGGGGG"}' mature.fa

[natasha321]

sorry, I understand the mature mirna sequences are small and it's difficult to generate a usable index, but, please, why 25 Gs? Why not 25 As or Cs? Why 25?

I've been reading some papers and most of them either map on the genome or on the precurssor miRNA. I'm using Bowtie and I would like to align on the mature miRNAs, but I'm I'm having some problems with that (I get 99% not aligned).

I was thinking about changing some parameter on the bowtie-build, but I also don't know what would help. Anyone has any tips?

thanks in advance!