[mira_talk] Re: clusterise big input

Bastien Chevreux a écrit :

On Montag 04 Mai 2009 Laurent MANCHON wrote:

I need to make contigs using mira2 but my input file contains 450 000 ESTs.
So, i need to use different percent of alignment like 100%, 90%, 80% and
generate contigs and ACE files.
Do you know if Mira is able to treat this big input and which options i
need to use to perform this task ?
The input size is no problem. However, MIRA is an assembler and not a clusterer. What you should do is: assemble with MIRA and take that output (contigs + singlets + debris) to a clusterer.
You *could* go by hand (or a wrapper), using multiple calls of MIRA in a fashion like this: assemble all at 100% (setting all the appropriate flags, there are a few I'm sorry). Then, taking out the contigs formed and using just remaining singlets and debris, assemble all at 90% in the next stage, etc.pp 
Regards,
  Bastien


okay thank you bastien.

this the command i use to assemble my 450 000 ESTs:
mira -project=BIGINPUT -job=denovo,est,normal,454 -SK:mnr=yes -SK:rt=4 -GE:not=2 -CO:asir=yes 454_SETTINGS -AL:mrs=90:egp=yes:egpl=reject_codongaps:megpp=100 -CO:rodirs=10 -AL:mo=60
do you see another parameter to add to my command line to adjust the results ? 

Laurent --

Other related posts: