Re: [AMOS-help] Assembly error correction using AMOS

[Adam P. <aph...@gm...>]

In my experience, assembly correction is a difficult problem. My usual
strategy is to simply break contigs at obvious assembly problems. Often
I'll run the assembly multiple times with different parameters or
assemblers and use the validation stats as a measure to pick the best
looking assembly.

To do anything other than breaking (i.e. patching or joining contigs)
requires full knowledge of the assembly graph at that position, otherwise
you chance doing more harm than good. So I prefer to just redo that
assembly, using the validation as a guide of what went wrong during the
last assembly.

Mike Schatz has also done some work on merging multiple assemblies
(metaassembler). This strategy uses multiple assemblies to patch a
reference assembly based on which assembly scores best at each position of
the genome.

Best,
-Adam


On Tue, Feb 19, 2013 at 12:03 PM, Nicolas Tourasse <nic...@ib...
> wrote:

> Hello Dear AMOS community members and genome assembly guys,
>
> I have a question about genome assembly correction with AMOS.
> We have assembled the genome of a unicellular green alga (~80 Mbp in size)
> using MSR-CA.
> We have then checked the assembly quality using AMOSvalidate and Hawkeye,
> and now we would like to correct the detected mis-assembly features, such
> as compressions and expansions. We are thinking of using the correction
> tool stitchContigs to fix these regions. Our plan would be to do it by
> writing a pipeline running stitchContigs and AMOSvalidate and updating the
> AMOS bank iteratively, which is not trivial.
> So, before going into this, we would like to ask you if you known whether
> there already exist an automated assembly correction pipeline, based on
> AMOS or other software. We could not find such a pipeline so far.
> If not, do you think that our plan would make sense or that it doesn't
> seem to be realistically feasible?
> Also, how to correct other types of mis-assembly features such as
> HIGH_LINKING_CVG (reads with mates in another scaffold - should we break
> the contigs/scaffolds)?
> Basically, in summary, what are assembly experts usually doing for
> correcting their assemblies? We are aware of tools for sequencing error
> correction and assembly quality check, but not really for assembly
> correction.
>
> We would be very grateful for any tips and tricks you may have, and we
> thank very much in advance,
> Sincerely,
>
> Nicolas Tourasse and Olivier Vallon.
> --
>
> Nicolas J. Tourasse, PhD
> Institut de Biologie Physico-Chimique (IBPC)
> CNRS UMR 7141/Université Pierre et Marie Curie (UPMC)
> 13 rue Pierre et Marie Curie
> 75005 Paris, France
> tel: (+33) (0)1 58 41 51 91
> fax: (+33) (0)1 58 41 50 22
> e-mail: nic...@ib...
> http://www.ibpc.fr/UMR7141/
>
>
> ------------------------------------------------------------------------------
> Everyone hates slow websites. So do we.
> Make your web apps faster with AppDynamics
> Download AppDynamics Lite for free today:
> http://p.sf.net/sfu/appdyn_d2d_feb
> _______________________________________________
> AMOS-help mailing list
> AMO...@li...
> https://lists.sourceforge.net/lists/listinfo/amos-help
>