This is a test version of Biostars. For the public version, visit https://www.biostars.org.
Extract all paired reads from sam file

Hi,

I had to filter manually a sam file by keeping only reads that fall into a specific regions.

Using flagstat, my sam file is looking like this now:

160995 + 0 in total (QC-passed reads + QC-failed reads)
1148 + 0 secondary
0 + 0 supplementary
0 + 0 duplicates
159950 + 0 mapped (99.35% : N/A)
159847 + 0 paired in sequencing
79922 + 0 read1
79925 + 0 read2
142206 + 0 properly paired (88.96% : N/A)
157757 + 0 with itself and mate mapped
1045 + 0 singletons (0.65% : N/A)
10605 + 0 with mate mapped to a different chr
6694 + 0 with mate mapped to a different chr (mapQ>=5)

As you can see, there is not an even number of reads and the number of read1 is different of read2.

I would like to extract only paired reads (and make a new sam file) whenever they are mapped or unmapped.

I have tried using the flag -f 1 but it gaves the same result.

Anybody have solution ?

Thanks for your help.

sam paired
• 2,665 views • link
updated by ATpoint ATpoint • written by Picasa Picasa

do you want to extract paired end reads that map to reference in to fastq files?

• 0 views • link
written by reza reza

1 answer

Probably by subsetting to a specific region, one of the mates was excluded, while the other one is still present and flagged as paired. Use fixmate to update the flags, then rerun the filter:

samtools sort -n -l 0 -O bam in.sam | samtools fixmate - - | samtools view -h -f 1 -o out.sam
• 0 views • link
written by ATpoint ATpoint

Thanks,

It looks like to work.

Just a quick question:

if a read is present and its mate has been excluded, with your command, does it remove the paired or reinclude the mate ?

• 0 views • link
written by Picasa Picasa

Exclude. Fixmate will flag the read as unpaired (this is why sorting by name is necessary, so fixmate can check if mates are both present) and -f1 will remove it.

• 0 views • link
written by ATpoint ATpoint

Log in to answer this question.

More posts like this