Cellucid — Interactive Single-Cell Data Visualization
Cellucid is a free, open-source, GPU-accelerated web app for interactive single-cell data visualization. It shows millions of cells in your browser, in 1D, 2D and 3D embeddings such as UMAP, colored by gene expression or cell metadata.
Key features
- Real-time rendering of millions of cells with adaptive level of detail (LOD)
- 1D, 2D and 3D embeddings such as UMAP, several per dataset, side by side
- Gene expression overlays with efficient sparse matrix handling
- Categorical and continuous cell metadata coloring, filtering, and selection
- Animated vector field overlays (RNA velocity / drift) with GPU particle flow
- Connectivity edge visualization (KNN graph)
- Publication export: SVG (vector) + PNG (high-DPI)
- Community annotation voting with optional GitHub sync
Data sources
Cellucid reads these files in the browser, in place, without uploading them: AnnData (.h5ad, Zarr store folders and .zarr.zip archives), MuData (.h5mu and MuData Zarr), Loom (.loom), 10x Cell Ranger HDF5 (.h5) and outs/ folders, MatrixMarket folders, and Seurat or SingleCellExperiment objects saved from R (.rds, .rda). Open one from your computer, a link, a dataset catalog or a server. The cellucid Python package (Scanpy/AnnData) and the cellucid R package (Seurat, SingleCellExperiment) prepare data for it and show data in it from a script, a Jupyter notebook or an R session.
Who it is for
Biologists and bioinformaticians who want to explore a single-cell atlas without installing anything: explore scRNA-seq atlases, query marker genes, compare embeddings side by side, visualize trajectories and RNA velocity, and build consensus community annotations.
Citing
Cellucid is free and open source under the BSD 3-Clause license. To cite it, use the reference on the Citing Cellucid page of the documentation.
