README_bam2gvcf_makeOarJobs

####################################
19/10/2023
NTM

GOAL:
run bam2gvcf_* on a cluster, where the queue system is OAR.


HOW:
In order to run bam2gvcf_* on a cluster such as f-dahu, we need working binaries
of the programs used by bam2gvcf_* , and functional installs of the perl
modules (deps).
On f-dahu we were only missing Parallel::ForkManager, installed with:
# clean out old stuff
rm -rf ~/perl5/*
cpan App::cpanminus
cpanm  Parallel::ForkManager

For gatk I rsync'd the latest SIF image, produced on our server luxor with:
singularity pull docker://broadinstitute/gatk:4.4.0.0

For strelka I gave up: it requires python 2, which isn't available on our compute nodes.
Anyways Strelka is fast enough to run on a simple server.


I also wrote 2_bam2gvcf_gatk_makeOarJobs.pl to submit 2_bam2gvcf_gatk.pl jobs via OAR,
$samplesPerJob samples at a time.
The script has some hard-coded stuff (paths of BAMs and path where GVCFs should go,
naming convention for samples == grexomeXXXX, oarsub command and parameters to submit
jobs on our compute cluster, etc...) but may still be useful.


EXAMPLE:
processing BAMs for samples grexome0751 -> grexome0916 with:
perl Bam2gvcf_GATK_Binaries/2_bam2gvcf_gatk_makeOarJobs.pl 751 916 &> bam2gvcf_gatk_submitJobs.log