SEQanswers (
-   Bioinformatics (
-   -   Qsub bwa pipeline... (

milesgr 06-27-2012 10:15 AM

Qsub bwa pipeline...
This may seem like a relatively easy question, but here goes:

I am trying to create a relatively simple shell script that sits in a parent folder, goes through all subfolders, grabs all files with the extension .fq, and aligns them using BWA aln. The caveat is that I need to use qsub to do this or I will make the admins very unhappy :)

My initial script was something along the lines of the following:

result=(`find . -name "*.fq" -type f`)

for i1 in ${result[@]}

qsub -q queue_name -l nodes=1:ppn=8 -V


It should be noted that the first portion simply gets the filenames (and paths) and the second calls, which is shown below (note that the first two variable paths are truncated for forum purposes):


result=(`find . -name "*.fq" -type f`)

for i1 in ${result[@]}

$BWA_HOME/bwa aln -t 8 $REFERENCE $i1 > $i1".sai"


I am simply trying to get BWA to run using the $1i filename and $i1.sai as the output name. The reason I ran the for loop here and in the initial script is because there seems to be a problem holding on to $i1. If I run the bottom script as a standalone script, BWA begins running on the first file. The key is I'm not certain how to get it to run qsub X times (where X is the number of .fq files), each time is with a different file/output. Any help would be greatly appreciated. Thanks!

milesgr 06-28-2012 10:44 AM

Anyone? :)

bjchen 06-28-2012 11:07 AM

What if you change your to just take one .fq file at a time and you pass the file name to in the for-loop of your initial script?

Jon_Keats 06-28-2012 08:09 PM

If you are interested check out our pipeline ( Its a qsub based program. We build two flavours. One that reads all the fastq pairs in a folder and aligns then calls variants. The other leverages an input file that allows for proper RG tag assignment that is required for GATK variant calling. Nice thing is it will merge multiple lanes of the same library together for you.

alec 06-29-2012 03:01 AM

You'll probably find it more convenient to use a single job array than submitting many jobs in a loop.
In your launch script do a single qsub:
qsub -t 0-$[${#result[@]}-1]
In the environment variable PBS_ARRAYID tells you which job is running:

milesgr 07-02-2012 08:33 AM

Got it working using:

sh "/filepath/ $i1" | qsub -q queue_name -l nodes=1:ppn=8

Not positive why I had to go about it in such a roundabout way.

All times are GMT -8. The time now is 10:57 PM.

Powered by vBulletin® Version 3.8.9
Copyright ©2000 - 2020, vBulletin Solutions, Inc.