-
Notifications
You must be signed in to change notification settings - Fork 0
/
model.job
33 lines (30 loc) · 941 Bytes
/
model.job
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
#!/bin/bash
#
#SBATCH --job-name=DITTO
#SBATCH --output=DITTO_logs.out
#
# Number of tasks needed for this job. Generally, used with MPI jobs
#SBATCH --ntasks=1
#SBATCH --partition=amd-hdr100-res
#SBATCH --time=06:00:00
#
# Number of CPUs allocated to each task.
#SBATCH --cpus-per-task=1
#
# Mimimum memory required per allocated CPU in MegaBytes.
#SBATCH --mem=10G
#
# Send mail to the email address when the job fails
#SBATCH --mail-type=FAIL
#Set your environment here
module reset
module load Java/13.0.2
module load Anaconda3
#conda activate nextflow
#Modify paths and run the pipeline here
/data/project/worthey_lab/tools/nextflow/nextflow-22.10.7/nextflow run ../pipeline.nf \
--outdir /data/results \
-work-dir .work_dir/ \
--build hg38 -c cheaha.config -with-report \
--sample_sheet .test_data/file_list.txt -resume
#https://training.nextflow.io/basic_training/cache_and_resume/#how-to-organize-in-silico-experiments