[mira_talk] Re: After Scaffolding
- From: Bastien Chevreux <bach@xxxxxxxxxxxx>
- To: mira_talk@xxxxxxxxxxxxx
- Date: Wed, 16 Sep 2009 22:50:41 +0200
On Freitag 11 September 2009 Davide Sassera wrote:
> 1 Repeats: could you please tell me what nnr would you set in this
> situation?
>
> 0 8200338
> 1 1253412
> 2 168676
> 3 30776
> 4 6770
> 5 3196
> 6 2272
> 7 2074
> 8 1538
> 9 568
> 10 476
> 11 208
> 12 338
> 13 306
> 14 318
> 15 258
> 16 210
> 17 166
> 18 212
> 19 196
> 20 200
> 21 198
> 22 234
> 23 152
> 24 100
> 25 72
> 26 28
> 27 26
> 28 32
> 29 42
> 30 44
> 31 32
> 32 42
> 33 54
> 34 68
> 35 32
> 36 64
> 37 86
> 38 40
> 39 78
> 40 68
> 41 50
> 42 56
> 43 48
> 44 50
> 45 48
> 46 26
> 47 46
> 48 32
> 49 30
> 50 22
> 51 12
> 52 4
> 56 2
> 57 6
If MIRA does not complain (MEGAHUBS and such), none. If it complains: try 20.
If still complaining: 15, 10, 9, 8, 7, ...
> 2 sequencing errors: you said that having more zeros than 1s is bad, so
> in my situation what should I do? jump off a cliff? change sequence
> company?
Jumping of cliffs higher than 2.57cm is highly discouraged as this might have
negative side-effects on the environment in case of blood and gore splashing
around. Save the planet, just don't do it :-)
Now, things are not always as easy as I write. There are cases when the repeat
histogram is not straightforward:
- misestimation of the repetitiveness of the data by SKIM. This happens often
with highly repetitive genomes (>20%-40%) and always with non-normalised
EST.
- high-coverage 454 and/or Solexa data
To dump the sequence company you'd have to look at the error rate of the reads
ion a finsihed project (telling MIRA *not* to use the sequence editor as this
would get you away quite a lot of them)
> 3 chimeras: I remember in a previous version that after the annoying
> request of some italian guy (me) you implemented a chimera finder. it
> found around 4000 in my assembly. Is it possible to set some parameter
> to make this search less lenient. I'm willing to lose some good reads in
> order to get read of chimeras
Nothing you can do as this is one of the rare algorithms in MIRA where one
can't tweak anything at the moment: either there are overlaps spanning reads
or not. What you can try is using higher number for -SK:bph, but I doubt that
it will lead to any changer.
Why do you want to tweak? Do you have the impressions that there are still
chimeras?
Regards,
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: