Location: /scratch.global/neis/bakeoff/
These scripts will submit RUFUS jobs (detailed below)
runAll : Submit a batch of samples for a group as separate jobs using the 1000G SNP exclusion list.
Usage: ./runAll <group, i.e. group1> [batchsize] [partition (default: msismall)]
runAll_1000G : Submit a batch of samples using the 1000G exclusion list
Usage: ./runAll_1000G <group, i.e. group1> [batchsize] [partition (default: msismall)]
runArray : Submit a batch of samples for a group as job arrays (use only if you're able to use amd2tb or amd512 or msilarge).
Usage: <group name> [num chunks (default: remaining samples)] [partition (default: amd2tb)]
rufus_docker_nonarray.slrm : Use for submitting a batch of samples as separate jobs.
rufus_docker_single.slrm : Use for running a single sample in a container.
rufus_docker.slrm : Use for submitting multiple batches of jobs with job arrays.
rufus_single.slrm : Use for submitting a single sample not in a container
backup_rufus.slrm : Back up all RUFUS output to tier2 storage. This runs automatically in runAll
If a job fails, usually re-running it does the trick. This can be done by just running runAll again normally, since it'll just run any incomplete samples.
If the same sample fails multiple times (the sample name and group are logged), it may help to run without a container:
sbatch -p <partition> --export=GROUP=[group name, i.e. group1],SAMPLE=[NWD###],EXCLUSION=1000G_SNP rufus_single.slrm
Single run:
sbatch -p [partition] --export=GROUP=[group],SAMPLE=[NWD###] cue_docker.slrm