[mira_talk] Re: some questions

Bastien,
it's not very easy to choose a valuable combination of parameters to obtain "good" contigs and i think i need to make many test adding or removing some parameters or changing values of them to actually 
find the best combination.
I just need to use mira to built good consensus sequences to then blastx them to swissprot or nr databases, this 
is the main goal of my project. So, assembling step is a very decisive step.
I think you should include in the manual of mira (in the futur)some specific examples(command line with parameters) that people who have already used in their experiment to treat particular problem. And people who have checked the best parameters choice used with mira to treat their problem should send you as example the command line they used in their specific context. 
I think it might be easier to create that and it's work a try.

Laurent --


Bastien Chevreux a écrit :

On Mittwoch 06 Mai 2009 Laurent MANCHON wrote:

thank you so much for this help Bastien,
just a question:
Do you see something wrong or not good in my command line below:

mira -project=ESTUR -job=denovo,est,normal,454 -SK:mnr=yes -SK:rt=4
-GE:not=2 -CO:asir=yes -CO:mr 454_SETTINGS
-AL:mrs=95:egp=yes:egpl=reject_codongaps:megpp=100 -CO:rodirs=10
-AL:mo=60 -CL:cpat
The definition of "wrong" depends entirely on your goals. You have a mixture of lenient and less lenient alignment parameters.
-CO:asir=yes will lead to "clustering" of close sequences, even if they'd "really" differ in vivo. 

-AL:mrs=95 is a very strict parameter, as is the combination -
AL:egp=yes:egpl=reject_codongaps:megpp=100

-CO:rodirs=10 agian relaxes things a bit.
I'd say that this command line leads to mixture of "true EST sequences" and some "clustering-like sequences". But that's just a guess. 

Regards,
  Bastien

Other related posts: