This script makes running multiple (BEAST)[http://beast.bio.ed.ac.uk/] jobs easier by looking
for a standard directory structure, with the BEAST XML file in the current directory and one
subdirectory for each replicate you want to run, and then submitting jobs to the cluster
for each subdirectory it finds. So if you are in /usr/people/myuser/mybeast
and the current directory contains mybeastjob.xml
the script looks for all
subdirectories (e.g. ones names rep1
, rep2
, etc) and starts a BEAST job
in each subdirectory using mybeastjob.xml
as input.
The default way of running runbeast.sh
is:
runbeast.sh <XML filename>
where <XML filename>
is the filename of your BEAST XML file. E.g.
runbeast.sh mybeastjob.xml
This uses the default version of BEAST installed on the cluster. To see which versions of BEAST are available use the command:
module avail beast
By default BEAGLE support is also enabled.
Further settings can be passed to the script using environment variables:
Variable Name | Purpose | Example |
---|---|---|
BEAST_HMEM | Amount of RAM requested for BEAST job | export BEAST_HMEM=10G |
BEAST_SEED | Manually set the BEAST random seed | export BEAST_SEED=1437498772784 |
BEAGLE | Flag passed to determine BEAGLE usage | export BEAGLE=-beagle |
BEAST_MEM | Amount of heap space allocated to java VM | export BEAST_MEM=4096m |
BEAST_STACK | Amount of stack space allocated to java VM | export BEAST_STACK=128m |
These need to be set before running runbeast.sh
.
Read the file LICENSE
for the licensing terms of the contents of this repository.