Running the Pipeline¶

Snakemake basics¶

All commands are issued from the repository root directory. The root Snakefile is the entry point; the -s flag should not be used.

snakemake --cores 4    # 4 parallel jobs (use --cores 1 for debugging)

To preview what will be executed without running anything:

snakemake -n --cores 4

Target	What it runs
`snakemake --cores 4`	Everything (default -- all scenarios, all stages)
`results/{folder}/{scenario}/epimight.done`	EPIMIGHT heritability estimation
`results/{folder}/{scenario}/scenario.done`	All stages for one scenario
`results/{folder}/{scenario}/simulate.done`	Pedigree simulation only
`results/{folder}/{scenario}/phenotype.done`	Simulation + phenotyping
`results/{folder}/{scenario}/validate.done`	Simulation + validation + folder summaries
`results/{folder}/{scenario}/stats.done`	Phenotyping + stats + plots

snakemake --cores 4 results/base/baseline10K/scenario.done

The scenario.done sentinel file signals that all stages are complete for that scenario.

Use -f to force-rebuild a specific output:

# Regenerate plots for a scenario
snakemake --cores 4 -f results/base/baseline10K/plots/atlas.pdf

The pipeline runs stages in order, with each stage depending on the previous:

simulate -> dropout -> phenotype -> censor -> sample -> stats/validate -> plots

Snakemake tracks file dependencies automatically; re-running the same command after an interrupted run resumes from where it stopped.

When Snakemake detects incomplete files from a previously interrupted run:

snakemake --cores 4 --rerun-incomplete

Problem	Solution
`ModuleNotFoundError: No module named 'simace'`	Run `conda activate simACE` first
`FileNotFoundError: config/_default.yaml`	Run snakemake from the simACE repo root directory
Simulation killed or frozen (large N)	Reduce `--cores` to lower parallel memory usage
`IncompleteFilesException` on re-run	Run with `--rerun-incomplete`