Cluster cells using the Leiden algorithm, the GPU-accelerated
rapids-singlecell implementation built on cuGraph.

t-SNE (t-distributed Stochastic Neighbor Embedding) is a non-linear
dimensionality reduction technique for visualizing high-dimensional data.
It computes a low-dimensional embedding from a precomputed representation
in .obsm (e.g.

UMAP (Uniform Manifold Approximation and Projection) is a manifold learning
technique for visualizing high-dimensional data.