reference-based contig assembly?

[lukas1848]

Hi all,

I am facing a task which should be really easy to tackle, but I am kind of stuck. Maybe one of you has an idea of how to solve it.

I have a set of contigs (ranging from 2 to 120 kb) coming from a single bacterial species. I would like to assemble these contigs but I know that they do not overlap, i.e. that some parts of the genome are not present in my contigs. I therefore want to map my contigs against a reference genome and then just call a consensus alternative genome from my contigs, where all the gaps are filled with Ns.

I already tried samtools together with vcfutils' vcf2fq but this just outputs the reference genome again.
Here's my command for calling the consensus

samtools view -uS sorted.sam |samtools mpileup -uf genome.fasta - | bcftools view -cg - | vcfutils.pl vcf2fq > consensus.fq

So, has anyone an idea which alternative tool I could use to do that?

I also thought about writing my own script and somehow create fake paired end contigs from the information I get mapping the contigs against the reference.

[winsettz]

Have you tried Mauve Contig Mover?

[GenoMax]

Find mauve here

[krobison]

MIRA assembler has a map-to-reference based assembly mode; AMOS suite also contains a comparative assembly tool.

Also, Scaffolding low quality genomes using orthologous protein sequences. It looks like GRASS might also work for you.

Probably many more

[winsettz]

And velvet's columbus mode, amongst others.

I'm curious if original poster assembled the reads naively (as in without reference) and is comparing against a reference? I did a short de novo test using ILM's DH10B MiSeq reads with and without the reference some time ago...

[yzzhang]

Also, there is a PAGIT pipeline published on nature protocol, This pipeline uses ABACAS to map contigs on reference, fill the gaps with Ns, then use the original reads to fill these gaps. Does this method work?