[mira_talk] Re: mapping

On Tuesday 14 April 2009 Giuseppe D'Auria wrote:
> Dear all, I noticed in my last mapping assembly that there are "things"
> assembled on my backbone genome where the name is something progressive
> ####1#### to ###nnnn###. The nucleotides are in lowercase and are
> immediately below the backbone. Always perfectly overlap the backbone.
> The strange thing is that sometimes overlap my reads sometimes not. I am
> quite sure they do not belong to my reads
> Somebody know what they are?

Yes. You shouldn't be seeing them at all ... or did you look at some 
intermediate files from the log directory? If they're in the final results, 
it's 
a bug.

These reads are "rail reads" (or "rails"), i.e., reads that MIRA creates 
internally during mapping assemblies to guide and speed up assembly. They also 
allow nifty things like almost perfect read distribution across different 
repeat sites.

Depending on your "--job", MIRA will use certain presets ... but these may not 
be ideal for Titanium reads or the newer Solexa 76mers. In case you need to 
tweak them: the parameters for these are in the -SB section. Basically, choose 
twice the size of your largest reads as length (-SB:brl) and the length as 
overlap (-SB:bro). If using technologies with different lengths, take the 
longer. E.g. for 454 FLX (assuming an average length of 250 bases) and Solexa 
76 mers: -SB:brl=500:bro=250


Regards,
  Bastien


-- 
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

Other related posts: