I was running Bowtie 2.2.3 using the -a option. In this mode, I expected to get mapping qualities of either 0 or 255. However, I was seeing MAPQ scores in the resultant SAM file of 14, 18, 21, 34, etc. I suspect this is a bug. Here's a snippet of the first five columns of a SAM file where this is happening: the middle four reads have MAPQ 34, while the others have MAPQ 255.
-bash-3.2$ samtools view -q 9 -S SG1_all_alignment.sam | cut -f 1,2,3,4,5 | head -n 248055 | tail -n 8
[samopen] SAM header is present: 1 sequences.
SEA055490C:246:C3LLPACXX:2:1211:4929:80295 99 gi|9626372|ref|NC_001422.1| 4453 255
SEA055490C:246:C3LLPACXX:2:1211:4929:80295 147 gi|9626372|ref|NC_001422.1| 4805 255
SEA055490C:246:C3LLPACXX:2:1211:10915:80467 83 gi|9626372|ref|NC_001422.1| 3748 34
SEA055490C:246:C3LLPACXX:2:1211:10915:80467 163 gi|9626372|ref|NC_001422.1| 3412 34
SEA055490C:246:C3LLPACXX:2:1211:10915:80467 339 gi|9626372|ref|NC_001422.1| 3748 34
SEA055490C:246:C3LLPACXX:2:1211:10915:80467 419 gi|9626372|ref|NC_001422.1| 3379 34
SEA055490C:246:C3LLPACXX:2:1211:5156:80671 83 gi|9626372|ref|NC_001422.1| 2436 255
SEA055490C:246:C3LLPACXX:2:1211:5156:80671 163 gi|9626372|ref|NC_001422.1| 2152 255
More generally, I'm using Bowtie2 to align reads to a reference that contains several genomes. I was comparing the alignments I get when I split this reference into two chunks, say, of 5 genomes each vs. aligning to a reference containing all 10 genomes. I was expecting that, after using the -a option, I would see the same number of reads aligned per reference genome regardless of whether I split the reference set or not. Instead, I did not see this result. Is this a bug?
Log in to post a comment.