Reference guided assembly

[StopCodon]

Hi,

What are some of the tools for reference guided assembly? I am trying to assemble a Eukaryotic genome around 800-1000MB in length. There is a closely related species genome available (about 70% of Illumina 100 bp reads from this genome align to the related species). I do not have a large RAM machine to do a true de novo assembly (my machine is about 24GB RAM). I have about 50x coverage from Illumina and SOLiD paired-end reads. MOSAIK seems to be one tool that claims to do this but I found zero citations to the MOSAIK publication for assembly.

[Adjuvant]

I'm going through the process of trying to figure this out right now as well. I'm working with a prokaryotic genome that is 15 - 20% variable from the available reference genomes. I've looked at MOSAIK several times and have come to the conclusion that it's a problem of word-usage. The program is referred to alternatingly in its own documentation as an aligner and an assembler, but functionally it still only aligns to reference like bwa or novoalign and their ilk. The options for true hybrid reference-guided assembly seem to be EXTREMELY limited. There's Velvet's Columbus module, which I've worked with a bit. I'm also in the middle of trying to work through the method outlined in Schneeberger et al's paper which seems like it may have potential to successfully combine the benefits of alignment and assembly, but is a really long multi-step manual process and relies on some older tools (AMOScmp, for example). I'm about halfway through a work-through, so I can't tell you whether it's any better than pure de-novo or Velvet Columbus yet, but take a look and see whether it has the potential to help you.

[mht]

Hi,

I was wondering if you've had any results from trying out both Velvet Columbus and the method from the paper? I'm looking to do some reference-based assembly and so would be nice to have this rough benchmarking information at hand.

Thanks!

[upendra_35]

Hi,

What are some of the tools for reference guided assembly? I am trying to assemble a Eukaryotic genome around 800-1000MB in length. There is a closely related species genome available (about 70% of Illumina 100 bp reads from this genome align to the related species). I do not have a large RAM machine to do a true de novo assembly (my machine is about 24GB RAM). I have about 50x coverage from Illumina and SOLiD paired-end reads. MOSAIK seems to be one tool that claims to do this but I found zero citations to the MOSAIK publication for assembly.

Why not try Tophat/cufflinks pipeline for assembling transcripts based on your reference. I am currently doing that with good results

[mht]

Hi upendra_35,

Any suggestions for genome assemblies? Not transcripts?

thx

[pari_89]

Hi upendra_35,

Any suggestions for genome assemblies? Not transcripts?

thx

Hi, Newbler performs reference-guided assembly. I ran it from the command line. It is runMapper option.

Hope this helps