FAIR Nextflow And AWS Primer

FAIR means data and workflows should be Findable, Accessible, Interoperable, and Reusable. A workflow does not become FAIR just because it runs; it needs metadata, provenance, versioning, and clear outputs.

flowchart LR
    A["Dataset manifest"] --> B["Nextflow workflow"]
    C["Containers and versions"] --> B
    D["Parameters"] --> B
    B --> E["Results"]
    B --> F["Provenance"]
    E --> G["Reusable report/app"]

FAIR Choices In This Project

• stable dataset accessions in data/dataset_manifest.yaml
• parameterized paths
• standardized results folders
• provenance JSON
• documented caveats and public-data limitations
• test profile for quick validation

AWS Adaptation

The awsbatch profile is a template for S3-backed execution with AWS Batch. Before using it, replace:

• YOUR_AWS_BATCH_QUEUE
• YOUR_ECR_OR_PUBLIC_CONTAINER
• s3://YOUR_BUCKET/...

Common Pitfalls

• Hardcoded local paths.
• Missing dataset versions or accessions.
• Reports without provenance.
• Cloud workflows that cannot run locally in a small test mode.