Extract fastq files of unaligned reads with Bowtie 2

[Mad4Seq]

Hi everyone
I am using Bowtie 2 (2.0.0-beta2) to do alignments on the output reads of an Illumina HiSeq 50bp paired-end RNA-seq experiment.

A preliminary analysis indicated that I have some rRNA condamination that is skewing my alignment quality metrics and I would like to get rid of those reads before further processing.

I have tried the --un option in two different ways:

1)As paired-end
bowtie2 -a -x /path1/path1/Scer_rRNA -1 /path2/path2/READS_R1.fastq -2 /path2/path2/READS_R2.fastq --un ./READS_NOrRNA.fastq

2)As single-end
bowtie2 -a -x /path1/path1/Scer_rRNA --un ./READS_NOrRNA.fastq -U /path2/path2/READS_R1.fastq -S /path2/path2/READS_not_R2.sam

(Also note the slight variations on the --un suntax)

In both cases I got the same error :Warning: Could not open read file "./READS_NOrRNA.fastq" for reading; skipping...

Does anyone have any ideas of how I should use the --un option or other ways to do the same thing?

Thank you.

[Mad4Seq]

Ok...
It seems like although the on-line manual of Bowtie2 lists --un <path> as an output option, it doesn't support this (see bowtie2 --help, it is not listed there). I used this option with Bowtie instead and seems to be working fine.

[Lcontami]

With Bowtie2 you should try to change the extension .fastq to .fq

[cookiemic]

Is this problem solved?
I encounter this problem too
I have tried fq instead of fastq
but still result an empty file

[cookiemic]

Thanks Kennels for solving my problem.

http://seqanswers.com/forums/showthread.php?t=31242