Tutorials

Executable notebooks mirroring the rapids-singlecell tutorials. Notebooks 1, 2, and 4 are self-contained (they auto-download their datasets); brain_1M needs the 10x 1.3M-neuron .h5 placed at data/external/1M_neurons.h5.

• 01 · Basic single-cell workflow on the Apple-silicon GPU
• 02 · Analytic Pearson residuals
• 04 · Spatial analysis (squidpy-GPU)
• 1.3 Million brain cells on a laptop GPU

Notebook	Workflow
01 · Basic single-cell workflow on the Apple-silicon GPU	QC → normalize → HVG → scale → PCA → neighbors → UMAP → Leiden → markers → Harmony → diffmap (PBMC 3k)
02 · Analytic Pearson residuals	Analytic Pearson-residual normalization → PCA → clustering (PBMC 3k)
04 · Spatial analysis (squidpy-GPU)	Spatial graph, Moran’s I / Geary’s C, co-occurrence (squidpy IMC)
1.3 Million brain cells on a laptop GPU	Full 1,000,000-cell workflow on a laptop GPU