Tutorials

Worked examples on a real dataset: endothelial cells (EC) from mouse spinal cord, comparing uninjured controls (UN, 3 replicates) against spinal cord injury (SCI, 3 replicates). See References & Data Sources for the full citation and GEO accession.

• Active Transcription Scoring on Real Spinal Cord Injury Data
• scATrans without spliced/unspliced layers: differential expression + enrichment + plotting

• Active Transcription Scoring on Real Spinal Cord Injury Data — for data with spliced/unspliced (or mature/nascent) layers: the full active_score workflow (heuristic, pseudobulk, permutation, gamma robustness, bias correction, advanced mode) end to end on real SCI vs. UN data.

• scATrans without spliced/unspliced layers: differential expression + enrichment + plotting — for data without spliced/unspliced layers: differential_expression across backends, plus a full tour of enrichment methods (ORA, KEGG, GO all-ontology, GSEA, redundancy reduction) and the plotting gallery.

Reproducing these notebooks locally

Both notebooks ship with their outputs already rendered, so you can read them here without running anything. To re-run them yourself:

git clone https://github.com/leelieber2025/scATrans.git
cd scATrans
pip install -e ".[dev,advanced,pseudobulk,gene_features,memento]" gseapy
jupyter lab docs/tutorials/t_ec_active_transcription.ipynb

EC.h5ad is included at the repository root, which is where both notebooks load it from (sc.read_h5ad("../../EC.h5ad")).