sc-grn

name	sc-grn
description	Infer gene regulatory networks from scRNA-seq using the pySCENIC workflow: GRNBoost2 for adjacency inference, cisTarget-style motif pruning, and AUCell regulon scoring.
version	0.2.0
author	OmicsClaw
license	MIT
tags	["singlecell","grn","pyscenic","regulon","transcription-factor"]
metadata	{"omicsclaw":{"domain":"singlecell","allowed_extra_flags":["--allow-simplified-grn","--cluster-key","--db","--motif","--n-jobs","--n-top-targets","--seed","--tf-list","--r-enhanced"],"param_hints":{"pyscenic_workflow":"[Truncated]"},"saves_h5ad":true,"requires_preprocessed":true,"requires":{"bins":"[Truncated]","env":"[Truncated]","config":"[Truncated]"},"emoji":"S","homepage":"https://github.com/OmicsClaw/OmicsClaw","os":["macos","linux"],"install":["[Truncated]"],"trigger_keywords":["grn","gene regulatory","scenic","pyscenic","regulon","transcription factor","grnboost"]}}

Single-Cell GRN

Why This Exists

• Without it: regulon analysis requires several external resources and multi-step orchestration.
• With it: pySCENIC-style GRN outputs are standardized into one wrapper contract.
• Why OmicsClaw: tables, figures, and AnnData outputs are bundled together.

Core Capabilities

1. Implementation-aligned pySCENIC workflow: adjacency inference, motif pruning, and regulon scoring.
2. Explicit external-resource contract: TF list, cisTarget databases, and motif annotations are surfaced as real prerequisites.
3. Regulon-centric outputs: adjacency, regulon, target, and AUCell tables.
4. Direct figure exports: regulon activity UMAP, regulon heatmap, and network plot.
5. Downstream-ready export: writes adata_with_grn.h5ad, report, result JSON, README, and notebook artifacts.

Scope Boundary

This skill currently exposes one public workflow: pyscenic_workflow.

That workflow covers:

1. GRNBoost2 adjacency inference
2. motif-based pruning
3. AUCell regulon scoring

Input Formats

Format	Extension / form	Current wrapper support	Notes
AnnData	.h5ad	preferred	most realistic path for preprocessed scRNA-seq
Shared-loader formats	.h5, .loom, .csv, .tsv, 10x directory	technically loadable	usually still need preprocessing and external GRN resources
Demo	--demo	yes	bundled lightweight fallback

Input Expectations

• The current skill expects a preprocessed scRNA-seq object.
• Required external resources: TF list, cisTarget database files, and motif annotations.
• The wrapper does not automatically fetch pySCENIC resources for the user.

Workflow

1. Load preprocessed expression data.
2. Run adjacency inference.
3. Prune candidate targets with motif evidence.
4. Score regulon activity per cell.