how to filter BWA paired-end alignments?

guru · 08-24-2011, 12:36 PM

Hi,

I want to filter my BAM alignments (generated using BWA on paired-end Illumina reads) to retain only those read pairs where at least one of the reads of the pair has unique mapping. Is there a way of doing this using the tag information available in the SAM file? or does anyone know of a tool that does something on these lines? Any suggestion will be greatly appreciated.

Thanks in advance..
Guru.

stefanoberri · 08-24-2011, 01:07 PM

I'd like to know this answer as well.

As far as I understand, if one of the two has uniq mapping, than the pair has unique mapping (maybe not, it could be that the second one can map in two very close places, quite unusual though). If the pair has unique mapping, than their score is > 0

However, this is my understanding, but I am not quite sure. If anybody has a better explanation, I am eager to listen...

swbarnes2 · 08-24-2011, 01:32 PM

bwa's sampe adds an XT tag to the sam file that indicates if the read is uniquely mapped, or repetitively mapped. So you could pull all of those reads out with awk, get the cluster coordinates, and then pull out all the reads from the .bam that had those coordinates.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Simulating paired-end reads and bowtie alignments	droog_22	Bioinformatics	0	02-09-2012 08:53 AM
bowtie- paired end - no alignments	madsaan	Bioinformatics	1	06-27-2011 02:24 PM
Support for parallelization of paired-end alignments with BWA	Fabien Campagne	Bioinformatics	0	12-17-2010 05:40 AM
Filter paired end BAM file based on iSize	Leif Bergsagel	Bioinformatics	2	12-16-2010 12:50 PM
Bowtie got different alignments for fliped paired-end	狗熊趴趴	Bioinformatics	0	06-26-2010 08:39 AM

08-24-2011, 12:36 PM	#1
guru Junior Member Location: State College, PA Join Date: Jul 2009 Posts: 1	how to filter BWA paired-end alignments? Hi, I want to filter my BAM alignments (generated using BWA on paired-end Illumina reads) to retain only those read pairs where at least one of the reads of the pair has unique mapping. Is there a way of doing this using the tag information available in the SAM file? or does anyone know of a tool that does something on these lines? Any suggestion will be greatly appreciated. Thanks in advance.. Guru.

08-24-2011, 01:07 PM	#2
stefanoberri Member Location: Cambridge area, UK Join Date: Jan 2010 Posts: 35	I'd like to know this answer as well. As far as I understand, if one of the two has uniq mapping, than the pair has unique mapping (maybe not, it could be that the second one can map in two very close places, quite unusual though). If the pair has unique mapping, than their score is > 0 However, this is my understanding, but I am not quite sure. If anybody has a better explanation, I am eager to listen...

08-24-2011, 01:32 PM	#3
swbarnes2 Senior Member Location: San Diego Join Date: May 2008 Posts: 912	bwa's sampe adds an XT tag to the sam file that indicates if the read is uniquely mapped, or repetitively mapped. So you could pull all of those reads out with awk, get the cluster coordinates, and then pull out all the reads from the .bam that had those coordinates.