Skip to content
Key Points

Aggregated resource

Key Points

This page is generated from episode key points.

6 episodes covered
24 takeaways
Best for recap and publish checks

Episode 1

Getting Started: Rules, Targets, and the DAG

5 takeaways collected from this episode.

  • A workflow is defined in a Snakefile.
  • Snakemake works backwards from the output you request.
  • Rules are connected by matching outputs to inputs.
  • rule all defines the default target for the workflow.
  • Snakemake only reruns outputs that are missing or stale.

Episode 2

Scaling with Wildcards for Parallel Processing

4 takeaways collected from this episode.

  • Wildcards let one rule match many input and output files.
  • Snakemake can run many independent jobs in parallel and then combine their outputs in a later step.
  • expand() is a convenient way to define a collection of target files.
  • script: is useful when a workflow step is more naturally written as a small Python script than as a shell one-liner.

Episode 3

Dynamic File Discovery with Checkpoints

4 takeaways collected from this episode.

  • Use a checkpoint when the downstream file list is only known after an earlier step has run.
  • A checkpoint lets Snakemake pause, run a discovery step, and then reevaluate the DAG.
  • Input functions can inspect checkpoint outputs and construct the downstream targets dynamically.
  • If the file list is already known in advance, ordinary wildcards and expand() are simpler.

Episode 4

Containers

4 takeaways collected from this episode.

  • container: lets a rule declare the software environment it needs.
  • --software-deployment-method apptainer tells Snakemake to execute containerised rules with Apptainer.
  • Per-rule containers keep workflow logic and software requirements explicit.
  • Some details depend on where you run the workflow, for example, whether Apptainer is already installed or whether extra bind mounts are needed.

Episode 5

Running on HTCondor

4 takeaways collected from this episode.

  • The Snakefile should usually stay the same across local and batch execution.
  • A workflow profile is the clean place to store HTCondor-specific executor settings.
  • Resource requests and batching strategy strongly affect queue efficiency.
  • The native HTCondor executor plugin avoids older submission wrappers.

Episode 6

Bonus: Visualising the Workflow

3 takeaways collected from this episode.

  • DAG: A visual map of your analysis dependencies.
  • Dry-run (-n): Always perform a dry-run to verify the plan before executing.
  • Rule Graph: A simplified visualization showing the relationship between rules rather than individual files.