Le 22/05/2014 21:13, Bastien Chevreux a écrit :
On 22 May 2014, at 17:25 , Laurent MANCHON <lmanchon@xxxxxxxxxxxxxx> wrote:no no sorry, it's a mistake, it was 454 single end. and the problem is to detect RNA splicing between 2 conditions. I don't know if it is possible to do that with Mira. My first idea was to use first gmap to map the reads onto genome, then run cufflinks on the resulting bam .Disclaimer: my experience with viral sequences is minimal. I once played around a little bit with public data, but that is just about all. That being said, maybe your problem can be reduced a bit. You wrote detected … does that mean “counting” of known variants or “discovery” of new variants?
just counting
For counting of known variants, what I’d do would be a simple test by simply putting all splicing variants
this is here my question, how to select or isolate this bag of splicing variants from mira results ?
into a reference file and map your reads against that (with relatively struct settings). Of course, reads coming from regions which are in common with all variants will be evenly distributed, but reads from certain splicing variants will automatically be mapped to the variant matching the best. Then it’s just a counting problem. You can even “improve” that detection by simply taking the splice junction +/- 30 bp as bait for “mirabait”. If the 454 data is from late 454 machines, there is some hope that not too many sequencing errors would be there and again you’d have a simple counting problem. If it’s for de-novo discovery … hmmm … maybe I’d try a mapping with strict settings followed by a de-novo of unmapped reads. At least for Illumina that’s work. B.
-- You have received this mail because you are subscribed to the mira_talk mailing list. For information on how to subscribe or unsubscribe, please visit http://www.chevreux.org/mira_mailinglists.html