- Install the runner framework and the CoNVex pipeline
# Set DIR to your installation directory, here it assumes your home directory
DIR=$HOME
cd $DIR
git clone git://github.com/VertebrateResequencing/vr-runner.git
git clone git@github.com:HurlesGroupSanger/CoNVex.git
export PERL5LIB="$DIR/vr-runner/modules:$PERL5LIB"
export PATH="$DIR/vr-runner/scripts:$DIR/CoNVex/utils:$PATH"
# Test if paths were set correctly
run-convex -h
# There are a number of default paths to files and executables in the pipeline.
# They can be overriden by providing a config file:
run-convex +sampleconf > convex.conf
# After editing the config file, provide the +config parameter to the pipeline,
# in addition to the options described below
run-convex +config convex.conf ...
-
How to run runner pipelines in general.
Best is to use the
screencommand:
# Open a new persistent screen called "canoes". The "-S convex" part is
# not required, but helps when multiple screens are left in background.
screen -S convex
# Do some terminal action here
...
# Now detach from the screen, leaving the terminal magically active in the background
Press `CTRL+a' then press `d'
# Later, attach back to this screen using
screen -r
# If multiple screens are running, choose the desired one by providing the name
screen -r convex
Documentation to the runner framework can be found here https://github.com/VertebrateResequencing/vr-runner/
- How to run the run-convex pipeline
# Create a template configuration file
run-convex +sampleconf > convex.conf
# Using your favorite text editor, set paths to the reference file and other resources
vi convex.conf
# Run the pipeline
run-convex -o outdir/ -b bam-sex.txt +config convex.conf +loop 300 +maxjobs 300 +mail pd3
The meaning of the options:
-o outdir
.. all the outputs will be placed here
-b bam-sex.txt
.. list of BAMs and sexes, for example
/lustre/scratch115/projects/ddd/rearchive_data/20150220/1866STDY5139686.bam M
/lustre/scratch115/projects/ddd/rearchive_data/20150220/1866STDY5139688.bam F
+loop 300
.. tell the pipeline to check periodically (every 5 minutes) the running jobs.
If the option is not given, the pipeline checks the running jobs, spawns
new jobs if necessary and exists. In this mode it can be run from crontab.
+maxjobs 300
.. to make things faster, don't spawn more jobs than this at once
+mail pd3
.. if the pipeline finishes or something goes wrong, email this user
- When making changes to the java code
cd inst/java/src
DIR=`realpath .`
export CLASSPATH=$DIR/../lib/sam-1.67.jar:$DIR/../lib/args4j-2.0.12.jar:$DIR/../lib/biojava-1.7.1.jar:$DIR/psicquic-simple-client-1.5.3.jar
mkdir -p rmme && cd rmme && javac -d . ../*.java && jar -cvf ../../lib/CoNVex.jar *.class; cd ..