bowtie2 --n-ceil parameter

[zamirabqi]

Hello everyone,

I'm trying to map some specific reads to a reference genome.
The problem is these reads has some "N" characters in them. Actually my mapping is from reads like this:

GCT (N)n CATCGGA

with n = 0 to 30 size.

I'm trying to use bowtie2 to make the alignments considering the Ns and no other mismatch with this:

bowtie2 -f -N 0 --np 0 --n-ceil 0,0.75 -x genome reads -S result.sam

My problem is to set --n-ceil parameter. I thought that the maximum Ns in my sequence is 30 in a 40 sequence length (75% of the sequence are N).

Does anyone here knows what is the problem? I just have to adjust this parameter?

Thank you.

[Brian Bushnell]

Reads that are 75% N sound like garbage; I wouldn't waste time with them. If you do want to use them you should quality-trim them, which will get rid of the Ns (assuming you use a trimmer implementing the phred algorithm, like BBDuk). But probably most of them will end up too short after trimming to map usefully.

[zamirabqi]

It's not a proprely read from a sequencing.

I'm making a test to search for targets. That's why I have to let this gap of Ns between the sequence.

[dpryan]

If you're just looking for the locations of a few motifs, then just use biopython or bioperl. The various short read aligners will often not work well when the read is filled with Ns.