Standardize and harmonize gene names
skillskillsetup L1★1
chansigit/stangene ↗What it does
Harmonize gene identifiers across single-cell transcriptomics datasets
Best for
Cross-dataset gene alignment in single-cell RNA-seq when feature naming standards differ.
Inputs
- · h5ad, .tsv, or .csv files with feature names; target species (human/mouse)
Outputs
- · harmonization_table.tsv, summary.json, conflicts.tsv, unmapped.tsv, optional *_harmonized.h5ad
Requires
- · stangene (Python package), Ollama references (downloads ~15MB human, ~7MB mouse)
Preconditions
- · stangene installed
- · References built for target species
- · Species specified explicitly
Failure modes
- · Auto-resolved ambiguities without user approval
- · Overwriting original identifiers
- · Cross-species harmonization without per-species passes
Trust signals
- · Conservative merge optional (policy: strict or symbol)
- · Never auto-resolves — surfaces conflicts for user decision