Skip to content

Repository files navigation

Install/dependencies:

dorado==0.5.1
snakemake==7.X.X
samtools

Runtime:

Pre-run:

cd /gpfs/commons/groups/innovation/dwm/basecall_snake
module load samtools
module load dorado

Run w/ slurm:

snakemake --cluster-config slurm_config.yml \
--cluster "sbatch --mail-type {cluster.mail-type} -p {cluster.partition} -t {cluster.time} --mem {cluster.mem} -D {cluster.chdir} --output={cluster.output} --gres={cluster.gres}" --cluster-cancel \
-j 4 -k --local-cores 8
snakemake --profile profiles/slurm -j 4 -k --local-cores 8

Note: in cluster execution -j controls the number of parallel gpu jobs that will be used to run basecalling, while --local-cores controls the number of local cpu threads that will be used for other tasks

Run w/out slurm:

snakemake --nt -k -j 2

Additional detiails:

Dorado models:

link to info from ONT How to download:

cd basecall_snake/
dorado download --directory resources/dorado models

Guppy models:

  • Location of all models on cluster:
    /nfs/sw/guppy/guppy-6.1.2-gpu/data/

Output tree (#TODO):

{OUTDIR}
    {sample}
        guppy
        dorado
            sup
                unaligned.bam
            hac
                unaligned.bam

TODO:

  • add snakemake v8 compatibility
  • Fix slurm resource settings for list_sample_runs
  • snakemake --lint
  • Autodetect chemistry?
  • Add other basecallers
  • README
    • output tree
    • model info

About

Snakemake workflow to run basecalling on ONT data

Resources

Stars

0 stars

Watchers

1 watching

Forks

Releases

Packages

Contributors

Languages