Confusion about the use of Bowtie software

[lix]

I am a new user of the Bowtie software.
For this software, there are so many option parameter settings, so which options should we usually consider or choose?
Thanks.

[SillyPoint]

Let's try a simple case first: a single-ended alignment using mostly default parameters.

First you need to convert your reference genome to a bowtie index -- for example:

cd ~/references
bowtie-build B-pertussis.fa B-pertussis

... starting with the .fa in your ~/references directory.

Then tell bowtie the directory where references reside:

export BOWTIE_INDEXES=~/references

Then run the alignment. You need to decide how many bases you want bowtie to consider. If you want it to use all the bases in the read, just do:

bowtie -v 3 --best B-pertussis reads.fastq

That will ask bowtie to give you the single best (fewest mismatches) alignment with <= 3 mismatches.

Alternatively, if you want to ignore some of the less-reliable bases towards the end of the read, you could do:

bowtie -l 32 -n 3 --best B-pertussis reads.fastq

(-l = first character of 'lemon'). That will cause bowtie to align using only the first (5'-end) 32 bases. Errors in the other bases will be reported, but they don't count towards the <= 3 mismatches.

The output produced by bowtie is documented in the 'bowtie --help' output. There is some confusion over forward/reverse strand reporting and numbering. What I found best there was just to stare at enough examples until I could make sense of what I was looking at.

Good luck.

--TS