Tutorials#
Browse tutorials#
Convert raw fragments into a Parquet-backed workflow, compute QC summaries, and inspect the features carried forward into downstream analysis.
Build a low-dimensional embedding, identify cell groups and annotate their cell types from gene activity, call peaks by cluster, and assemble a peak matrix for follow-up analyses.
Test cluster-specific peak sets for transcription factor motif enrichment and inspect the strongest motif signals in accessible regions.
Compute chromVAR deviation scores from motif annotations and visualize transcription factor activity across cells and clusters.
Rank marker peaks, derive motif-linked gene sets, and run enrichment analysis to summarize regulatory programs by cluster.
Correct Tn5 bias and score footprints per cell type, then test transcription factor binding between cell types with the GPU port of TOBIAS.
Analysis notebooks for GATAC are maintained in the companion
gatac-tutorials repository
and are included here as a git submodule at tutorials/ in the repo root.
Notebooks are rendered with MyST-NB — you can view them here or download and run them locally after installing GATAC.
Setting up the notebooks#
The notebooks are stored in a separate repository and linked as a git
submodule under tutorials/ at the repo root. To initialise:
git submodule update --init --recursive
To run notebooks locally:
# Install GATAC with full dependencies
uv sync --extra rapids-cu12
# Run a notebook
uv run jupyter lab tutorials/01_fragment_preprocessing.ipynb
Note
Pre-computed outputs are committed to the repository so the docs build does
not require a GPU. Set nb_execution_mode = "auto" in docs/conf.py to
re-execute notebooks during the docs build.