brainreg_sherlock

Some simple applications to run brainreg on Stanford's Sherlock SLURM cluster.

1. The data_sync utility copies files to scratch space for faster data access. This runs commands from mpifileutils.

2. The downsample utility downsamples the input data to 10/25/50um for registration to atlases at those resolutions. The brainreg app can do downsampling, but it is more efficient to do it once up front if brainreg is being called many times. This is a python utility that mimicks what brainreg does for downsampling.

3. The brainreg utility runs registration with a given set of parameters. It is run from a virtual environment housed in GROUP_HOME.

The files with _launch in their name are the top-level programs to run many sbatch commands, each of which starts a SLURM job. These can be modified for the location of your data. The actual sbatch scripts with the command lines are the {app}.sh files.