Bowtie output format question

[fahim]

Hi
I recently downloaded Bowtie and tying to align affy sequences. I am still using BLAT for longer sequences, but for short sequences, Bowtie is having clear advantage.

I have a doubt about the default output of bowtie. Field number FOUR(4) gives 0-based offset of the start location of reads on the genome. But it is not showing the location where mapping ends. For example, the BLAT output tells you everything including start, end , length of reads matched, mismatched etc. etc.

What I am assuming that since bowtie does not allow gapped alignment, the end should be equal to ( start+ lenghtOfReads-1). Is it true for the alignment results?
Thanks
Fahim

[xinwu]

Hi
I recently downloaded Bowtie and tying to align affy sequences. I am still using BLAT for longer sequences, but for short sequences, Bowtie is having clear advantage.

I have a doubt about the default output of bowtie. Field number FOUR(4) gives 0-based offset of the start location of reads on the genome. But it is not showing the location where mapping ends. For example, the BLAT output tells you everything including start, end , length of reads matched, mismatched etc. etc.

What I am assuming that since bowtie does not allow gapped alignment, the end should be equal to ( start+ lenghtOfReads-1). Is it true for the alignment results?
Thanks
Fahim

You have the answer and should consider the strand.

[kjlee]

From the bowtie manual:

0-based offset into the forward reference strand where leftmost character of the alignment occurs.

Leftmost of which sequence: the reference or the query? I believe that it does matter. I presume it is the leftmost of the reference. If this is the case, then the 5' end of the query would be at (offset+length-1).

[swbarnes2]

From the bowtie manual:



Leftmost of which sequence: the reference or the query? I believe that it does matter. I presume it is the leftmost of the reference. If this is the case, then the 5' end of the query would be at (offset+length-1).

The reference. But check the CIGAR string to see if the read has been hard or soft trimmed.

[kjlee]

Say the query is 36nts long. The sequence aligns to the negative strand and the last (most 3') nucleotides (let's say 5nts) are soft-clipped, will the position reported be the first position that has a properly-mapped (non-soft-clipped) nucleotide, or will it be the 3' most nucleotide of the 36?