Bowtie2 output with paired end reads

[sampie]

Hi,

I am trying to use bowtie2 to align paired end Illumina read files.

I am using following command:

./bowtie2 -a --threads 8 --n-ceil 42 --very-sensitive-local --quiet --no-sq --no-hd --np 0 --ma 1 --mp 0,0 --rdg 1,1 --rfg 1,1 --score-min L,0,0.95 -x rdp/bt2_index/rdp.db -q -1 sR1.fastq -2 sR2.fastq > bt2out.txt

The index is made from RDP II database and currently sR1.fastq and sR2.fastq contain only one read:

sR1.fastq:
@1_M01363:14:000000000-A5K8J:1:1101:8617:1211 1:N:0:16
TACGGAGGATCCGAGCGTTATCCGGA...

sR2.fastq:
@2_M01363:14:000000000-A5K8J:1:1101:8617:1211 2:N:0:16
TTCACCGTTGCCGGCGTACTCCCCAGGTGGA

The bt2out.txt file shows mostly results with sR1.fastq read header.

1_M01363:14:000000000-A5K8J:1:1101:8617:1211....
1_M01363:14:000000000-A5K8J:1:1101:8617:1211....
1_M01363:14:000000000-A5K8J:1:1101:8617:1211....
.
.
.

I am not sure how to read bowtie2 results. How can I obtain information how each matching read aligned as a pair?

Thanks

[westerman]

Bowtie2 should be outputting BAM/SAM files. Read up on that format. If you have further questions after reading about SAM files -- especially the bitwise flag field -- then ask a question.

[GenoMax]

It appears that sampie omitted "-S filename_to_hold_alignments.sam" option on the command line to write the alignments to a file.

[westerman]

It appears that sampie omitted "-S filename_to_hold_alignments.sam" option on the command line to write the alignments to a file.

While good practice the '-S' is optional. As per the docs:

-S <hit>
File to write SAM alignments to. By default, alignments are written to the "standard out" or "stdout" filehandle (i.e. the console).

So the stdout file of bt2out.txt should be a SAM file.

[jwfoley]

As long as we're talking best practices, I would definitely pipe the output through samtools and save a BAM rather than a SAM file. There's no reason a SAM file should ever exist on a hard drive.

[sampie]

Thanks for replies.

I took a look at SAM format specification. It was more clear to me than bowtie2 manual.

I am now testing bits 0x1, 0x2 and 0x8 from flags and if they are on I am determining that current match is R2 and previous match is R1 (of the same pair).

Hopefully it works correctly now.