A question on samtools rmdup

[wangzkai]

Hi,

I was studying the SAM protocol on SourceForge:

http://sourceforge.net/apps/mediawiki/samtools/index.php?title=SAM_protocol

There is one command that merge sorted alignments and remove potential PCR/optical duplicates:

samtools merge -rh rg.txt - ga.bam 454.bam | samtools rmdup - - | samtools rmdup -s - aln.bam

My question is why there are two "rmdup" operations? Is this a typo?

Thanks!

[nilshomer]

Hi,

I was studying the SAM protocol on SourceForge:

http://sourceforge.net/apps/mediawiki/samtools/index.php?title=SAM_protocol

There is one command that merge sorted alignments and remove potential PCR/optical duplicates:

samtools merge -rh rg.txt - ga.bam 454.bam | samtools rmdup - - | samtools rmdup -s - aln.bam

My question is why there are two "rmdup" operations? Is this a typo?

Thanks!

One for paired end, one for single end reads. Nevertheless, you should not use samtools for duplicate removal; use picard instead.

[wangzkai]

Thank you, nilshomer. On the samtools manual page, there are two rmdup commands:

samtools rmdup <input.srt.bam> <out.bam>

and

samtools rmdupse <input.srt.bam> <out.bam>

So does this mean that:

"samtools -s" is equivalent to "samtools rmdupse"?

I will check out picard. Is there a reason that you prefer it over samtools?

[nilshomer]

Thank you, nilshomer. On the samtools manual page, there are two rmdup commands:

samtools rmdup <input.srt.bam> <out.bam>

and

samtools rmdupse <input.srt.bam> <out.bam>

So does this mean that:

"samtools -s" is equivalent to "samtools rmdupse"?

I will check out picard. Is there a reason that you prefer it over samtools?

Since samtools does not remove duplicates properly (I am tool lazy to find the link to the author himself stating this).