← SkillSafe / Script Checker for deepTools

Check a deepTools script before you run it

Paste a ChIP-seq, ATAC-seq or RNA-seq coverage script, or generate one from the deeptools skill's templates. Your browser runs every command through an emulation of deepTools 3.5.6's own argument parser and checks normalization, read extension, genome size and Tn5 shifting, free. Nothing is uploaded. A paid run reviews the script or drafts a pipeline from your description, and the page checks that script too.

The examples replay saved model runs - the whole page, free.

1. The experiment

2. The script (bash - deepTools commands are checked, the rest is read for variables)

Start from the skill's template (workflow_generator.py, in your browser)
Empty fields use the script's defaults. The genome size default is the GRCh38 value.
script.sh
Drop a .sh file here, or
A regions BED to check too (optional - validate_files.py's BED check)
Paste a script, generate one, or load an example.
Check a script first to price the run.

Your recent runs

What this does, and what it does not

Each deepTools command stops at its first argument error before it reads any data. The page emulates Python 3.12's argparse over the argument declarations of all 20 deepTools 3.5.6 commands - read from the installed package, and matched against the real parser on thousands of generated command lines - so it reports the same first error, then the checks each tool's main() makes on its arguments alone (RPGC without --effectiveGenomeSize, bamCompare's normalization rules, missing outputs, label counts). It cannot open your files, so output paths and matrix files are not checked, and values from variables the script never sets are left alone.

The page's own checks, labelled "page", quote the deeptools agent skill and the deepTools docs: read extension for ChIP-seq and never for RNA-seq, duplicates after GC correction, the Tn5 shift for ATAC-seq, one normalization and one effective genome size per study, and a genome size that belongs to another assembly. One thing they found: the skill's own chipseq_qc template calls plotCorrelation without its required --whatToPlot, so it stops at that step. Derived from the agent skill @k-dense-ai/deeptools (k-dense-ai/scientific-agent-skills, MIT; see the notice). deepTools is by Ramírez et al. (Nucleic Acids Research, 2016). Independent tool, not affiliated with or endorsed by the deepTools project or the Max Planck Institute of Immunobiology and Epigenetics.