how to extract unique hits from a sam file

[gfmgfm]

I aligned reads with bwa and I want to get a set of the reads that mapped uniquely to the genome.
I understood from samtools faq that they suggest to look at 'reliable' rather than `unique' by :

samtools view -bq 1 aln.bam > aln-reliable.bam

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

However, I am interested to get the subset of the uniquely mapped reads, in order to do some calculations on it.

How can one do it?

[drio]

Take a look to this thread.

I'd suggest you follow bwa's FAQ advice and rely more in the MAPQ. Still, if you want to
filter uniquely mapped:

$ samtools view bwa.bam | grep "XT:A:U"

[gfmgfm]

Thanks a lot!

[seq_GA]

Hi,
I am also looking for such solution. But in my bam file, I don't see any

XT:A:U

instead its a SOLEXA single read where I have converted export to sam format as below.

test_1:8:69:19633:9434       0       chr10   5423197 4       45M     *       0       0       CACACAACCCCCACACCAAACACACACCCCCCACACACAACAAAC      0.2B90+)*=@8@################################   XD:Z:6C11C15G4CACT2     SM:i:4
test_1:8:56:11474:20981      0       chr10   7323903 6       45M     *       0       0       ATCAAGCGATCCTCCCACCTCATCCCCCTAAGTACCTGTGACTAA      757@54;3;1@@@@@22@###########################   XD:Z:22G11G3G5C SM:i:6

Any generic way of picking unique hits from sam file? Thanks.

[gfmgfm]

I am not an expert, but don't think you can extract this from the _export file. I think XT:A:U is specific to bwa (maybe also other aligners?).

Maybe you would like to use the _sorted file from the Illumina pipeline:
the desciption of the _sorted file- from CASAVA 1.7 manual p. 73:
This output file is similar to s_N_export.txt, except it contains
only entries for reads which pass purity filtering and have a
unique alignment in the reference. These are sorted by order
of their alignment position, which is meant to facilitate the
extraction of ranges of reads for purposes of visualization or
SNP calling.
These files are only produced if the flag WITH_SORTED is
used."

Alternatively, you can take your reads and align them with bwa (or with other aligner that gives you this info).

[emp]

hello all,
I have Illumina data, which I mapped through Bowtie2 and got a sam file.

Now I need to extract the reads which are being shown uniquely one time.

Kindly guide if something could be done for that.

I have tried a bit of commands for the above, but failed to get those reads separated.

Kindly guide ASAP.

[Wallysb01]

hello all,
I have Illumina data, which I mapped through Bowtie2 and got a sam file.

Now I need to extract the reads which are being shown uniquely one time.

Kindly guide if something could be done for that.

I have tried a bit of commands for the above, but failed to get those reads separated.

Kindly guide ASAP.

How about just use bowtie with the -m 1 option set?

[emp]

hello Wallysb01,

after searching a lot for Bowtie2 to get uniquely matched reads, I think bowtie 1 is the only way out.

thanku for the same.

[dpryan]

Unique alignments in bowtie2 have MAPQ>=2, so you can just filter the results by that.