I'm mapping a set of paired end Illumina reads against a reference with Bowtie2.

bowtie2 -x reference -1 reads_1p.fastq -2 reads_2p.fastq -S aligned.sam -t --met-file metrics --no-unal -p 10

The fastq files have the seq_id line in the format:

@HWI-ST731_27:5:1110:2607:21757#4@0/1

with the /1 or /2 at the end indicating the paired read number.

The Bowtie2 output file (original sam and sorted bam) lists the reads id without the trailing /1 or /2 for most of the alignments but not all.

Why is that? Can I change tha behavior without changing the alignment and reporting parameters?

It seems to interfere with pysam.AlignmentFile.mate(), because I get an error indicating that a given read has no mate when this is not the fact. What does happen is that one mate has the mate number removed while the other not.

>samtools view aligned_s2.bam | grep 'HWI-ST539_132:4:2108:14057:10095'
HWI-ST539_132:4:2108:14057:10095#2@0/2  163 chr_2   1937    0   98M =   2052    146 TTGAGCATGAATGGGCATATGGCTGGATAAATAAGACTGGTAATCATCCTATGAACATAATCGTGATTAAGAGATAGAAATATGATTAGAAAGTAGGA  ?@BDB??B>B<CDDHIJJJEACFHGIGIBHIHGBGGGEHG*?F?DFGF?DD99BDEG>FHBBBFGGIJJJGECHA>=A?;CFEC;CC;;ACEDCCCA@  AS:i:-40    XS:i:-40    XN:i:3  XM:i:11 XO:i:0  XG:i:0  NM:i:11 MD:Z:0N0N0N4G0T8A4A17C0C4A2A48  YS:i:-14    YT:Z:CP
HWI-ST539_132:4:2108:14057:10095#2@0    83  chr_2   2052    0   31M =   1937    -146    GGGTAGAAAGGTAGTTGGTGAGACAAACCAG :*>GFFED?HFBDAHE;A@DFHFDDDBDD@@ AS:i:-14    XS:i:-4 XN:i:0  XM:i:3  XO:i:0  XG:i:0  NM:i:3  MD:Z:0A12T9T7   YS:i:-40    YT:Z:CP

What is the meaning of the modification of the id of just one of the mates?

Thanks!