Graph-based clustering of cells on the sNN graph

This function will apply Leiden clustering on the sNN graph with the given resolution and add a column to the obs table.

Usage

find_clusters_sc(
  object,
  cluster_algorithm = c("leiden", "louvain"),
  res = 1,
  name = "leiden_clustering",
  modality = c("rna", "adt", "wnn"),
  seed = 42L
)

Arguments

object

SingleCells, MetaCells (or potentially other) class.

cluster_algorithm

String. One of c("leiden", "louvain").

res

Numeric. The resolution parameter for igraph::cluster_leiden() or igraph::cluster_louvain().

name

String. The name to add to the obs table in the DuckDB.

modality

String. On which modality to run the UMAP. One of c("rna", "adt", "wnn"). The two latter options are only available for multi-modal versions with the added data.

seed

Integer. For reproducibility.

Value

The object with added clustering in the obs table.

Examples

# Leiden on the cached sNN graph
sc <- demo_single_cells()
sc <- find_clusters_sc(sc, res = 1.0, name = "clusters")
table(get_sc_obs(sc)$clusters)
#> 
#>   0   1   2 
#> 169 166 165 

unlink(sc@dir_data, recursive = TRUE, force = TRUE)