samtools: splitting a bam file putting all scaffolds together

Question

[Fedster]

I have some bam files that I want to split. The reference genome has 22 chromosomes and a bunch of unplaced scaffolds. I can easily split the bam files chromosome by chromososme

samtools view my.bam chromsosmeI -b -o my_chromosomeI.bam

I can do the same scaffold by scaffold, but I would like to do extract all scaffolds with some sort of wild card (there are a lot of them) and have them all in the same bam output. The wild card part is the one I'm having issues.

[Pierre Lindenbaum]

how about using the option '-L' of samtools view ?

 -L FILE  only include reads overlapping this BED FILE [null]

[michael.ante]

I'd try to extract the chromosomes and scaffolds by samtools view -Hand create a simple bed-file out of it. Your scaffold, you'd like to group store in separate files which are then used as e.g. samtools view my.bam -L scffod_group1.bed - b -o my_scaffold_group1.bam .

Additionally, I'd add the -h parameter in the samtools view calls to include the header.

Answer 1

[ATpoint]

samtools idxstats in.bam | cut -f1 | grep <grepForScaffoldNames> | xargs samtools view -o scaffolds.bam in.bam

[Fedster]

works, if you add it as an answer I will accept it as the correct one. Cheers

Answer 2

[finswimmer]

Make sure you have your bam file indexed by samtools index. Then you can do something like this:

$ samtools idxstats EQ18-NGS_S4-final.bam|cut -f1|grep -v "*"|parallel 'samtools view -o {}.bam my.bam {}'

samtools idxstats gives a tab-delimited file with informations about mapped read on each contig. The first column contains the name, so we extract it with cut -f1. We need to exclude the row with the name * by grep -v "*" as this is the reserved name for unmapped reads.

Now we have a list of all contigs and pass this to parallel to start samtools view with contig name as region parameter.

fin swimmer

[Fedster]

I do not want to run all chroms and contigs in parallel, I want to get all the fragments mapping to any contig (not in a chromosome) out in one single step and dump in one single file. My problem is not listing all possible fragments/contigs/chroms, it using a wildcard in samtools view to get out stuff on an unmapper contig, no matter the contig.

[Fedster]

basically samtools view my.bam contig* -b -o output.bam does not work