Re: [wgs-assembler-users] 454 paired-end + wgs assembly

[Santiago R. <san...@gm...>]

Hi, Brian.

Thanks for the tips.

I completely forgot to mention the organism: I'm trying to assemble a
Xanthomonas axonopodis bacteria. It has a 5.2Mb genome with a ~65% GC
content and two plasmids (64Kb and 33Kb). The problem is that despite being
different in size, the plasmids are very similar because they share similar
genes. They share, not consecutively, almost 20-25Kbp.

I have two sets of data: a) the pool of plasmids sequeced single-ended, and
b) the whole genome (including the plasmids) sequenced paired-ended.

Though I managed to obtain a good number of scaffolds (18) using Newbler
v2.6 I was unable to close the plasmids. That's why I'm now trying with
Celera and what I'm trying to accomplish.

So, any further advice would be really helpful.

Best,

Santino


On Thu, Apr 5, 2012 at 2:39 PM, Walenz, Brian <bw...@jc...> wrote:

>  Hi, Santino-
>
> Without knowing what you’re assembling, I’d say you should put nothing in
> the spec file.  The biggest knob to change is utgErrorRate (default 0.03).
>  Lower can separate repeats, or fail to assemble anything.  Higher can
> overcome polymorphism, or over collapse repeats and duplications.  The
> second biggest knob would be unitigger=bog vs unitigger=bogart.  Bogart is
> still ‘experimental’ in the 7.0 release; better in CVS tip.
>
> You might get better compute performance by setting the various
> “concurrency” options (especially cnsConcurrency) to the number of CPUs you
> have, and tuning overlapper for your hardware.  Generally, the defaults
> work well for something this size.
>
> b
> --
> Brian Walenz
> Senior Software Engineer
> J. Craig Venter Institute
>
>
>
> On 4/5/12 1:14 PM, "Santiago Revale" <san...@gm...> wrote:
>
> Hello everybody!
>
> I'm new to paired-ends assembly, so I'm hoping you could help me.
> I have two 454 FLX Titanium libraries: one single-end (common wgs) and one
> pair-end (8Kb).
>
> I ran sffToCA two times, one for each library, using "-trim chop" for the
> pair-end library (amongst other paired-end parameters).
>
> Now, what parameters you suggest should I put in the specFile to perform
> this assembly for better results?
>
> Thank you very much in advance.
>
> Best,
>
> Santino
>
>
>