User Guide
In-depth documentation for each stage of the GeneCircuitry pipeline, from QC through reporting.
| Page | What you’ll find |
|---|---|
| Pipeline Overview | PipelineController architecture, step names, CLI flags, parallel execution, checkpoint integration |
| Preprocessing & QC | Cell/gene filtering, normalization, HVG selection, dimensionality reduction, clustering |
| GRN Inference (CellOracle) | Oracle object creation, PCA, KNN imputation, link inference |
| Gene Modules (Hotspot) | Autocorrelation testing, module detection, enrichment-annotated heatmaps |
| Plotting System | genecircuitry/plotting/ subpackage — QC, GRN, and Hotspot canonical plot functions |
| Reporting | HTML and PDF report generation with generate_report() |
| Stratified Analysis | Per-cluster parallel analysis, output layout, configuration |
| Checkpoints & Resume | How .checkpoint files work, re-running steps, clearing state |
Pipeline at a glance
AnnData (.h5ad)
│
▼
[Preprocessing & QC] QC filtering · normalization · HVG selection
│
▼
[Dimensionality Reduction] PCA · UMAP · Leiden clustering
│
├──────────────────────────┬──────────────────────────┐
▼ ▼ │
[CellOracle GRN] [Hotspot Modules] │
per-cluster TF networks autocorrelated gene modules │
│ │ │
└──────────────────────────┘ │
│ │
▼ │
[GRN Deep Analysis] NetworkX plots, scoring │
│ │
▼ │
[Reporting] HTML · PDF report │
│
◄─────────────────────────────────────────────────────┘
Optional: stratify by cell type, run each in parallel
All stages are optional and can be run selectively using --steps or --skip-* flags.