[mira_talk] Re: Incremental assembly

On Monday 29 December 2008 18:48, mark.rose@xxxxxxxxxxxx wrote:
> Is it possible and advisable to do incremental assemblies of new
> sequence with CAF or other files from a previous assembly or should you
> re-assemble using all available reads?

Sorry for the delay in responding, I've been out for a bit longer over 
Christmas.

Depending on what you want/expect, it is possible (or not). I know a number of 
users who used that in formar times to decide when to stop sequencing.

You would need to use a mapping assembly and specify that you still want new 
contigs to be formed. That would be something like this:

  mira ... --job=mapping,... -SB:abnc=yes 

The following will happen:

1) MIRA loads the backbone sequence(s)
2) MIRA loads the "normal" (in this case: new) sequences
3) the new sequences are first mapped (if possible) to the existing backbones
4) sequences that cannot be mapped are used to form new contigs

Note that you might want to play around with different stringency parameters 
to prevent the "new" reads to be mapped to very similar stretches in your 
backbone sequences.

Caution: MIRA wil NOT join contigs! I.e., if contig1 and contig2 could be 
joined by a number of "new" reads, this will not happen. The reads are likely 
to be added to one of the contigs), in case where the overlap is too small 
(smaller than -AL:mo) they would even form a new contig. One can then quickly 
join contigs manually (e.g. in gap4).

In case you want MIRA to do contig joining, you'd need to perform a full 
re-assembly (sorry). If you already edited what you had up to now, you could 
still take advantage of that cleaning by converting the cleaned project to 
CAF and use that as input.

Hope that helps,
  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: