[mira_talk] Re: Parsing SSAHA output with Mira 2.9.35x1

On Friday 05 December 2008 21:35, Gregory Harhay wrote:
> Hi Bastien:
> [...]
> Is another way to check to see if the ssaha clips are actually getting into
> mira? 

In a separate directory, assemble the three identical artificial sequences in 
the attachement with 

mira -project=sal_BM -job=denovo,est,accurate,454 -fasta -notraceinfo 
-noclipping 
454_SETTINGS -CL:msvs=yes > & log.txt   

then look up the resulting consensus. The clips given in the SSAHA file (and 
represented in lower case in the sequences) are gone.


> The reason why I discovered the bug is because I'm generating many megahubs
> that don't go away when I apply the ssaha clips.

Have a look at http://www.freelists.org/post/mira_talk/mmhr-problem,1 where 
the lower part of the mail gives a rough idea on how to find out what is 
wreaking havoc. I need to smooth out the whole thing, but it should give you 
an idea.

Looking at the fact that you are assembling ESTs (a non-normalised 
library?) ... I'd guess that some housekeeping genes may give you some 
headache.

Regards,
  Bastien

Other related posts: