scaffolding ABySS output using Bambus

bye

unread,

Oct 4, 2010, 1:34:51 PM10/4/10

to ABySS

Hi,

I'm new to ABySS and Bambus. Now I have an ABySS output contigs.fa
file de novo assembly of paired-end Illumina data, I want to use
Bambus to do scaffolding which require .contig and .mate file. I tried
abyss2ace, somehow it took for ever to run and gave me with a .ace
file with only one line. Is there any other tools that can convert
contigs.fa to .contig while keep the pairing informations(so I can
easily make the .mate file) as well?

Thanks a lot!

Shaun Jackman

unread,

Oct 4, 2010, 2:58:53 PM10/4/10

to bye, Rong She, ABySS

Hi Robin,

Rong She (cc'ed) has used Bambus with ABySS. Perhaps she can answer your
question.

Cheers,
Shaun

Rong She

unread,

Oct 4, 2010, 4:57:31 PM10/4/10

to ABySS

Rong She wrote:
> You need to generate the reads-to-contigs alignments first (I used
> MrsFast, but you can use bwa or bowtie). Then convert the resulted sam
> file to .contig file. I have a script that processes MrsFast sam output
> to .contig format. Note that was intended for parsing an older-version
> of MrsFast output (v2.2.0.1) and may not be exactly applicable to other
> aligners or newer version of MrsFast. Feel free to modify it to your needs.
> Rong

MFsam2contig1

Francesco Vezzi

unread,

Oct 5, 2010, 2:31:09 AM10/5/10

to Rong She, ABySS

Hi Rong
If I understand well ABySS final contig file contains the scaffolds,
hence a scaffolding phase is already done by ABySS itself. What
advantages could one take from reiterating another scaffolding phase
with Bambus?

Best
Francesco

--

Francesco Vezzi, PhD Student
Tel: +39 3805151541
Department of Mathematics and Computer Science,
University of Udine, Udine, Italy
&
Applied Genomics Institute (IGA), Udine, Italy
http://www.appliedgenomics.org

Rong She

unread,

Oct 5, 2010, 4:27:33 AM10/5/10

to Francesco Vezzi, ABySS

ABySS does de novo assembly. Its scaffolding is based on read pairs. Bambus can be used as a more general scaffolding tool that allows other evidences, for example, reference genome.

Just a word of caution though, the current version of Bambus (v2.33) may have problems handling moderate sized data. I tried it with a genome of about 100M and it crashed, which looked like a program bug. I was told it is going through some rewriting but not sure when there will be a next release.
Rong

________________________________________
From: Francesco Vezzi [ve...@appliedgenomics.org]
Sent: Monday, October 04, 2010 11:30 PM
To: Rong She
Cc: ABySS
Subject: Re: scaffolding ABySS output using Bambus

Hi Rong
If I understand well ABySS final contig file contains the scaffolds,
hence a scaffolding phase is already done by ABySS itself. What
advantages could one take from reiterating another scaffolding phase
with Bambus?