[mira_talk] Re: Mira assembly with reference
- From: Andrei Tudor <andrei.tudor@xxxxxxxxxxxxxx>
- To: "mira_talk@xxxxxxxxxxxxx" <mira_talk@xxxxxxxxxxxxx>
- Date: Fri, 15 Apr 2011 17:34:44 -0700 (PDT)
Thank you Bastien,
This exactly what i wanted, only a guide to de-novo assembly with a very close
strain to the one sequenced. But if mapping does not do that, what is mapping??
Also is there any assembler that does what i am searching for?
Cheers,
Andrei
________________________________
From: Bastien Chevreux <bach@xxxxxxxxxxxx>
To: mira_talk@xxxxxxxxxxxxx
Sent: Friday, April 15, 2011 5:45:55 PM
Subject: Re: [mira_talk] Re: Mira assembly with reference
On Friday 15 April 2011 10:45:07 Peter wrote:
> On Thu, Apr 14, 2011 at 9:41 PM, Andrei Tudor
> > I am wondering if it is possible to assemble some 454 reads, but using a
> > reference. I want to see if it will give better contigs.
> Yes, you can give MIRA a backbone sequence in de novo mode.
While it is possible, it is probably not what Andrei wants.
Loading a backbone in de-novo still performs a mapping assembly ... but with
de-novo parametrisation. Not sure this will work as expected.
Andrei: what you probably ask is whether MIRA can assemble some reads de-novo
using a given sequence as guide on how to build together and guide contigs.
Right?
If that was your question: no, this is not possible.
When given a backbone (reference sequence) MIRA will alway first perform a
mapping assembly using your 454 reads. If you ask nicely and use the right
incantations, MIRA will also assemble de-novo all reads it could not map. The
command line for that is (in principle):
mira --project=... --job=mapping,genome,454,accurate
-AS:nop=5 -SB:abnc=yes
...
and other options at will. Beware: instead of a mapping assembly (quite quick),
this has the time and memory foorptrint of de-novo (can be quite long).
B.
Other related posts:
