Fig. 1
From: Nested Stochastic Block Models applied to the analysis of single cell data
Modeling randomised data. A UMAP embedding of 3k PBMC after standard processing with Leiden approach (left) or nSBM (right). Cell grouping is consistent for both the approaches (Adjusted Rand Index \(ARI=0.869\)). B UMAP embedding of the same data after randomisation of kNN graph edges. In this case schist does not return any cell grouping, while optimisation of modularity finds up to 24 different cell groups