reference based assembly from bam?

[lukas1848]

Hi,

I have mapped 100 bp Illumina reads to my reference genome and I wondered if there is any tool that can create an assembly of my reads from the bam file I created or if I have to start with the fastq files again using a reference based assembler.

Maybe I am missing something, but shouldn't it be fairly easy to assembly scaffolds from bam files?

[swbarnes2]

Hi,

I have mapped 100 bp Illumina reads to my reference genome and I wondered if there is any tool that can create an assembly of my reads from the bam file I created or if I have to start with the fastq files again using a reference based assembler.

Maybe I am missing something, but shouldn't it be fairly easy to assembly scaffolds from bam files?

Velvet will take .bams as input, but it will treat them like fastqs. It won't use alignment data in the assembly, just the reads.

[Wallysb01]

There are a few programs that do reference guided genome assembly if that's what you're looking for. I'm not as familiar with these tools, but a quick google search found a few: Mosaik, which supports bams: http://bioinformatics.bc.edu/marthlab/Mosaik

Also AMOScmp (or AMOScmp-shortread): http://sourceforge.net/apps/mediawik...?title=AMOScmp

I know there are others, but again, this isn't something I do.

What are you trying to do anyway? Are you reassembling the genome a specific individual for calling variants, or is this transcriptome? Are you using a related species, or is this same species?

[lukas1848]

Thanks.
I sequenced individuals of a different population than the one used for the reference genome. The two populations have distinct phenotypes and I primarily want to call SNPs. But since I have the aligned reads, I thought it would be nice and easy to assembly these into scaffolds as well.

I found this pipe

samtools mpileup -uf ref.fa aln.bam | bcftools view -cg - | vcfutils.pl vcf2fq > cns.fq

from http://samtools.sourceforge.net/mpileup.shtml

and I wondered if the consensus sequence created here, is just what I was looking for. If not, what does the consensus sequence provide?

[Wallysb01]

Thanks.
I sequenced individuals of a different population than the one used for the reference genome. The two populations have distinct phenotypes and I primarily want to call SNPs. But since I have the aligned reads, I thought it would be nice and easy to assembly these into scaffolds as well.

I found this pipe

from http://samtools.sourceforge.net/mpileup.shtml

and I wondered if the consensus sequence created here, is just what I was looking for. If not, what does the consensus sequence provide?

Well, it will recreate a genome for that individual, I believe filling in any gaps based on the reference, though I'm not sure about that. So, it just depends on what you want.

[lukas1848]

I believe filling in any gaps based on the reference, though I'm not sure about that.

The consensus sequence would then basically conceal any larger deletions or insertions if gaps would be filled with reference sequence and insertions would just end up in the bin as unmapped reads, right?

So I guess to end up with a proper assembly for my sample, I would need to start a reference based assembly from scratch using something like Velvet.

[Wallysb01]

The consensus sequence would then basically conceal any larger deletions or insertions if gaps would be filled with reference sequence and insertions would just end up in the bin as unmapped reads, right?

So I guess to end up with a proper assembly for my sample, I would need to start a reference based assembly from scratch using something like Velvet.

I'm guessing if it sees big gaps in aligned reads it will not assume that it is actually a large indel, so yes, de novo assembly or reference guided assembly is probably your better option if you're looking for larger structural variation.