SOLiD denovo: contigs corrupted after conversion from double-encoding to nucleotide

[Alex8]

Hello.

I use SOLiD de novo pipeline (http://solidsoftwaretools.com/gf/project/denovo/) to assemble bacterial genome.
I use Mummerplot for proof-mapping the resulting contigs to reference genome. Double-encoded contigs (contigs.de) look OK. However, after they are converted to base-space by ASiD utility, they break apart into small pieces, see figures. Single contig is double-encoded, while multiple fragments are what it turns into in base-space.

Anybody has a suggestion on how to cure this problem?

PS: I tried older version of SOLiD pipeline (denovo_preprocessor_solid + velvet + denovo_postprocessor_solid + denovoadp) and the base-space contigs were uncorrupted. However, it is much slower, so I'd still like to fix the latest denovo tool.

[pepperoni]

Dear Alex8,
did you solve the problem? I want to assemble color space with velvet and don't know how to do it.
thanks

[Alex8]

Hi,
I downloaded denovo2.2 version which seems to have the bug fixed.
however, solid software tools site is retired now - if you don't find it anywhere else, can send you the distr.

[colindaven]

I stumbled across SOPRA recently, which might be helpful for you guys.

It contains the denovo scripts from Solid Software Tools as well.

http://www.physics.rutgers.edu/~anirvans/SOPRA/

[cement_head]

CLC Genomics Workbench has a two week free trial...

[colindaven]

I'm not convinced yet CLC is the best solution for de novo assembly of colour space reads. Sure, it's easy, fast and it maps reads back to contigs to attempt to eliminate indel/colour errors....
but most contigs seemed were under 1000bp.

It appears that the CS Velvet is capable of building longer contigs > 1000bp.

(Solid 5500xl, 75bp single end reads)

I can't quantify yet but will do when we get results back for a range of settings.

[pepperoni]

Are you using the base space fastq data provided by the 5500? or the color space csfasta? otherwise how are you being able to use Columbus? Since my data came from SOLiD 4 I don't have any base space data.
thanks