changing bam files with mixed sample data to per sample

[mathew]

I have five Bam files which has data of two to three samples togather as the samples were run in same channel. Is there any easiest way of seprating Bam file for each samples i.e can I split Bam files and combine per samples.
Thanks.

[JackieBadger]

I have just started this kind o analysis and as far as I could figure out people have suggested aligning all of your samples separately and denoting each with a unique group header, then merging your alignments.

[mathew]

Bam per sample

JackieBadger, Thanks.
The problem which I have is each bam file is having two to three samples in it. So that means alignments are alreday performed? Any suggestion please how I can get it around to get bam per sample?

[qtrinh]

Hi,
You should be able to use samtools to split your bam file into multiple bam files.
Here is how:

# get the list of all sample names in INPUT.bam
> samtools view -H INPUT.bam | awk '$0 ~ /^@RG/' | cut -f 2 | sed 's/ID://' > sample.list
# for each sample 's' in sample.list; extract reads belong to this sample and save it as a bam file
>for s in `cat sample.list ` ; do echo $s > sample.txt ; samtools view -bo ${s}.bam -hR sample.txt INPUT.bam ; done

Hope this helps,

Q

[mathew]

Bam per sample

Thanks qtrinh,