Quickstart¶

Prepare the project¶

Place one assay in an experiment directory. Raw reads belong in FASTQ/:

experiment/
└── FASTQ/
    ├── sample_A_L001_R1_001.fastq.gz
    ├── sample_A_L001_R2_001.fastq.gz
    ├── sample_B_L001_R1_001.fastq.gz
    └── sample_B_L001_R2_001.fastq.gz

Use one consistent field separator in filenames. Metadata is recommended for any grouping or differential analysis.

Validate¶

micromamba activate omnomnomics
omnomnomics rna -i /path/to/experiment -g GRCh38 --dry-run

Replace rna with atac or chip as appropriate. A dry run validates inputs, resolves configuration, and builds the Snakemake DAG without submitting work.

Run¶

omnomnomics rna \
  -i /path/to/experiment \
  -g GRCh38 \
  -j all \
  --retention-policy pruned \
  --max-project-size 300G

The command submits a controller job with sbatch. The controller then submits the rule jobs, so do not submit the command through a second sbatch wrapper.

Monitor¶

omnomnomics monitor -i /path/to/experiment

The monitor reports workflow step states and recent pipeline log lines. Rule-level errors are written under slurm_logs/<rule>/.

Resume or rerun¶

By default, Snakemake reuses valid outputs. Select a later stage range to resume:

omnomnomics rna -i /path/to/experiment -g GRCh38 -j 5-12

Use --rerun-selected-steps only when those selected outputs must be rebuilt. It is not required after an ordinary interrupted run.