starCAT Web Application

A privacy-preserving web application for running starCAT (star-CellAnnoTator) directly in your browser.

Overview

starCAT annotates single-cell RNA-seq data with predefined gene expression programs and infers derived discrete annotations such as cell type labels. This web implementation uses Pyodide (Python compiled to WebAssembly) to run the complete starCAT algorithm entirely in your browser.

Key Features

• Complete Privacy: All computation happens client-side - your data never leaves your machine
• No Installation Required: Run sophisticated bioinformatics analysis directly in your browser
• Multiple Input Formats: Takes raw count data in both H5AD (AnnData) and MTX (Matrix Market) file formats
• Curated Reference Catalogs: Access to pre-built reference datasets for various tissues and species
• Zero Infrastructure: No backend servers or API limits

Quick Start

1. Visit the application (https://immunogenomics.io/starcat)
2. Select a reference catalog from the dropdown menu
3. Choose your file format (H5AD or MTX)
4. Upload your data:
   • H5AD: Single .h5ad file (up to ~50,000 cells or 700MB)
   • MTX: Three files (matrix, features, barcodes)
5. Click "Run starCAT" and wait for processing
6. Download results: Two TSV files (usage scores and detailed scores) plus interactive visualizations

Supported File Formats

H5AD Format

Standard AnnData format for single-cell data stored in HDF5 files.

MTX Format

Matrix Market sparse matrix format with three components:

• .mtx - Sparse count matrix
• features.tsv - Gene/feature names
• barcodes.tsv - Cell barcodes

Reference Catalogs

Pre-built reference catalogs are available for the following:

• Human T cells across tissies and diseases (TCAT.V1)
• Human myeloid cells from gliomas (MYELOID.GLIOMA.V1)
• Human bone marrow derived CD34+ hematapoetic stem cells (BONEMARROW.CD34POS.HSPC.V1)

View the full catalog at the References page.

Local Development

Since this is a static website, you can run it locally with any web server:

# Using Python 3
python -m http.server 8000

# Using Node.js http-server
npx http-server

# Or simply open starcat/index.html in your browser

Citation

If you use starCAT in your research, please cite:

Kotliar, D.*, Curtis, M.*, Agnew, R. et al. Reproducible single-cell annotation of programs underlying T cell subsets, activation states and functions. Nat Methods 22, 1964–1980 (2025). https://doi.org/10.1038/s41592-025-02793-1

Related Resources

• GitHub for starCAT algorithm
• Reference Catalogs: Available on Zenodo
• Documentation: See manuscript for methodology details