Remove header from BAM file

[tahamasoodi]

Hi,

How can we remove header from a BAM file?

[kmcarr]

Hi,

How can we remove header from a BAM file?

Do you mean to strip the header off entirely? You can't. Without a minimal header it is no longer a valid BAM file.

[swbarnes2]

If you really want to do that...

samtools view -b header.bam > noheader.bam

[kmcarr]

If you really want to do that...

samtools view -b header.bam > noheader.bam

That won't do it. I assume you are thinking that without the -h option samtools view will not output a header, but the -h option only applies to SAM (text) output. Your command will output exactly the same file as the input.

I think you were giong for

samtools view header.bam | samtools view -Sb - > noheader.bam

This command will give you a warning that no @SQ lines are present but will generate an output BAM. The resulting BAM will be worthless though. Without the @SQ information all of your alignment entries will have no chromosome name (just '*') and all of the positions will be '0'. You have essentially just stripped out the most important bits of data from the file.

[maubp]

If you want to edit or remove the embedded SAM plain text header in a BAM file, use the 'samtools reheader' command.

[tahamasoodi]

Thanks, if I replace the header will it have any effect on the chromosome name and positions?

[dpryan]

Thanks, if I replace the header will it have any effect on the chromosome name and positions?

It can. In BAM files, chromosome (or contig) names are only stored in the header. In the actual reads, what is stored is a numeric index saying where in the header the chromosome/contig name is located. So, if you shuffled around the chromosome ordering in the header without changing all of the reads, the reads would then be mapped to different places (same position, just different chromosomes). For that reason, one should reheader BAM files with care.

[tahamasoodi]

But the picard's ReplaceSamHeader will replace it in all the reads, then how come it affects the bam file?

[dpryan]

But the picard's ReplaceSamHeader will replace it in all the reads, then how come it affects the bam file?

As its name implies, the picard ReplaceSamHeader command is intended for SAM files, which are just regular human readable text files. In that case, there's no need to monkey with the reads. My comment is more in regards to the samtools reheader command, which quite literally just produces a BAM file with a different header.

[tahamasoodi]

By the way read ReplaceSamHeader thoroughly, it is not for SAM files only but it can be used for BAM files also. The original post was not related to samtools reheader but something different.

[swbarnes2]

That won't do it. I assume you are thinking that without the -h option samtools view will not output a header, but the -h option only applies to SAM (text) output. Your command will output exactly the same file as the input.

I think you were giong for

samtools view header.bam | samtools view -Sb - > noheader.bam

Ah, thanks. I was not at a computer where I could check it, and it's not anything I've actually wanted to do.