bowtie -m option

[frymor]

Hi everybody,

maybe I misunderstood it completely, but what is the option -m good for?
If I am looking for differentially regulated genes in my data set and i set a high m-value (lets say 5), I know there are for some reads up to five different possible locations which they can be matched.

@what exactly does it mean for my data?

Do I have my read at different positions on the reference genome?
This way my data for differentially expressed genes will be biased.
Do I have these reads at five locations or there is just one, best location where it mapped to.

I have pair-end data set, so I can't use the best and strata combination.
@What I don't exactly understand is the definition of uniqueness in the bowtie -m option?

If I use the option -m 1 I will loose a large part of my data, so I don't want to do it, but I would like to see, that the reads are mapped are as specific as possible.
the command I use to run bowtie is as such:

bowtie -a -m 5 -n 2 -l 22 -q --un total_trimmed.unmapped -t -p 2 -5 11 --chunkmbs 256 --max total_5trimmed.maxHits -S d_melanogaster_fb5_32 -1 s2_1_sequence.fq -2 s2_2_sequence.fq total_5trimmed.sam

[waseem]

m 1, as I understood will give you the reads that map to only 1 location. (unique reads). My approach is, these reads are counted as true counts (given more weights). The reads mapping at multiple locations might be separated into reads mapping at alternatively spliced variants (different variants of same gene models) and the reads mapping to different genes (multi reads). I am working on separating these reads and counting them with different weights.