Bowtie - What am I missing here?

[quantrix]

Ok so I am a newbie here, so please bear with me. I am trying out Bowtie. And following the manual and I issue the command as it is in the manual (I just removed the --suppress 1,5,6,7 to make all the columns visible.

./bowtie -a -v 2 e_coli -c ATGCATCATGCGCCAT

The first two lines of the output are as follows...

0 - gi|110640213|ref|NC_008253.1| 148810 ATGGCGCATGATGCAT IIIIIIIIIIIIIIII 0 10:A>G,13:C>G
0 - gi|110640213|ref|NC_008253.1| 2852852 ATGGCGCATGATGCAT IIIIIIIIIIIIIIII 0 8:T>A

If I am reading this right, it found matches on the negative reference strand, with TWO mismatches on the first result (10:A>G,13:C>G). But when I look at the alignments closely,

Test - ATGCATCATGCGCCAT
Result 1 - ATGGCGCATGATGCAT

I see mismatches at position 4,5,6,11,12,13.

Same thing with result TWO

Test - ATGCATCATGCGCCAT
Result 1 - ATGGCGCATGATGCAT

Bowtie says ONE mismatch in position 8, while I see mismatches in positions 4,5,6,11,12,13

What AM I missing here? Please apologize for relative ignorance of the question.
Thanks
Quantrix

[kopi-o]

If the match is to the negative strand, you probably need to take the reverse complement of one of your sequences. With that being said, it looks to me like your "Test" and "Result 1" are in fact identical taking the reverse complement, so there would be no mismatches rather than two.

[quantrix]

Dear Kopi,
You are right, in fact, I did the reverse complement for BOTH the sequences and they are exactly the same. So does this mean when Bowtie outputs the results, it DOES NOT output the actual result sequences? (And merely outputs the TEST sequence?)

I do not know. I am more confused than ever. Can someone clarify this for me please?
Thank you
Q

[kopi-o]

OK, I should have thought of that. What you are seeing is default Bowtie output, which according to the manual (http://bowtie-bio.sourceforge.net/ma...-bowtie-output) does not show the reference sequence, but the read sequence.

[quantrix]

Damnit!
RTFM...R-T-F-M! Now I feel like an idiot.
Thank you Kopi. But it is still weird though. Why even bother outputting the READ sequence when there is a match? I am sure there is an answer to it. But never mind.
Q

[ttnguyen]

All of the aligners I know output the read sequence (or reverse-complemented if it is mapped to reverse strand). This is helpful for the post-processes, e.g. you can convert the output back to fastq or you can calculate the GC content of the mapped reads.

[quantrix]

Hi Nguyen,
Thanks for the reply. It DOES make sense now. I think I got thrown off by the fact there was only ONE read in the test sequence. If there are millions of reads, it makes sense to output the reads and for other purposes such as GC content etc
Q