snake-basket

Description

This project provides multiples customizable rules for the workflow manager Snakemake. These rules have built as python function and can be imported (with snakemake include) and parametrized (with argument of the function).

Usage

In this paragraph we will add a rule markDuplicates in your workflow.

1. Import the rule

• Copy the rule in your workflow rules folder

   workflow_folder/
      rules/
          markDuplicates.smk          
      Snakemake
  

• Add the rule in the list of imports rules/all.smk

  Add all.smk in rules (once by workflow)

   workflow_folder/
      rules/
          all.smk
          markDuplicates.smk
      Snakemake
  

  Add the rule in all.smk (it contains all the rules to import)

   ...
   include: "markDuplicates.smk"
   ...
  

• Import all the wrapper functions in your Snakefile

   ...
   include: "rules/all.smk"
   ...
  

The function markDuplicates() is now accessibe in your workflow.

2. Call rule in your workflow

Add the function call and parameters in your code:

    ...
    markDuplicates(
        params_stringency="STRICT"
        params_keep_outputs=True
    )
    ...

All accessible parameters and their default values are presented in function declaration of the rule in markDuplicates.smk. The main categories of these parameters are: * input_, * output_, * param_, * snake_: for parameters related to the Snakemake element like wildcards restrictions. Keep in mind that input and ouput must be consistent in terms of wildcards like with a standard rule.

3. Set execution environment for the rule

You can provide software of the rule by one of this three ways:

• The software folder is in $PATH.

• The path of the software is in workflow configuration file:

    ...
    software_paths:
        picard: /home/user/bin/picard
    ...
  

  The name of the parameter is in bin_path argument of the params section of the rule in the markDuplicates.smk.

• You use conda environment with Snakemake (see option --use-conda) and the environment of your rule is in envs folder:

    workflow_folder/
        envs/
            picard.yml
        rules/
            all.smk
            markDuplicates.smk
        Snakemake
  

  The name of the environment file can be found in conda section of the rule in the markDuplicates.smk.

4. Configure the computing requirements [optional]

Optional in local and required with cluster submission.

Default value are stored in rule in resources section:

resources:
    extra = "",            # options added to cluster submission (example: "--qos=project_A")
    mem = "8G",            # maximum memory required
    partition = "normal"   # partition/queue name for job
threads: 1                 # threads per job

Declare parameter usage in snakemake command line:

snakemake \
  ...
  --cluster 'sbatch {resources.extra} --mem={resources.mem} --partition={resources.partition} --cpus-per-task={threads}'

Change default value by command line [recommended]

Values can be changed by --set-threads and --set-resources in snakemake command line.

snakemake \
  ...
  --cluster 'sbatch {resources.extra} --mem={resources.mem} --partition={resources.partition} --cpus-per-task={threads}' \
  --set-threads $RULE=$NUM \
  --set-resources $RULE:$KEY=$VAL $RULE:$KEY=$VAL \

Change default value by profile

You can change default values in profile configuration file.

$PROFILE/config.yaml
    cluster: sbatch
    cluster: "sbatch {resources.extra} --mem={resources.mem} --partition={resources.partition} --cpus-per-task={threads}"
    set-resources:
        - $RULE:$KEY=$VAL

Copyright

2019 Laboratoire d'Anatomo-Cytopathologie du CHU Toulouse