Hecatomb Snakemake for HTCF

1. Login to HTCF.
2. Move to your /scratch/ directory.
3. Clone the repo:

git clone --recurse-submodules https://github.com/RachelRodgers/hecatomb_htcf_snake.git

4. Edit the config.yaml file in the config directory to use your email in case of job failure:

--mail-user=yourEmailAddress

5. Make a directory to hold the snakemake profile:

mkdir -p ~/.config/snakemake/slurm_hecatomb

6. Copy the cluster submit and profile files to the appropriate locations:

cd hecatomb_htcf_snake
cp config/config.yaml ~/.config/snakemake/slurm_hecatomb
cp slurm-submit/*.py ~/.config/snakemake

7. Create a directory to hold your raw sequencing reads and move your data to into that directory.

8. Edit the hecatomb_config.yaml file (under /config/) to point to your data directory (under Paths: Reads) and edit the Read1 Read2 and Extension patterns as needed (under Patterns:). Note if your files contain both _R1.fastq.gz and _R1_L001.fastq.gz style designators, "_R1" is sufficient to capture both. Input files extensions can be .fastq or .fastq.gz. Uncompressed input files will be gzipped before the pipeline starts.

9. Submit in one of two ways:

   a. With sbatch script (preferred):

   sbatch submit_hecatomb_snake.sbatch
   

   b. Interactively (better for troubleshooting):

   # start an interactive session:
   srun --mem=48G --cpus-per-task=8 -J hecatomb -p interactive --constraint=cpu_E52650 --pty /bin/bash -l
   
   # load snakemake:
   ml snakemake/5.10.0-python-3.6.5
   
   # dry run (prints steps and stops):
   snakemake --profile slurm_hecatomb -n
   
   # production run (run steps):
   snakemake --profile slurm_hecatomb
   

10. See slurm output files in the logs_slurm/ directory which will generate inside the hecatomb_htcf_snake/ directory.