Re-runs Leiden community detection with new parameters and refreshes all
clustering-dependent downstream slots (cell-level IDs, composition matrices,
traditional fits). This is the recommended way to explore different
resolutions after the initial get.graph() call.
Usage
# S3 method for class 'Seurat'
recluster(x, ...)
# S3 method for class 'SingleCellExperiment'
recluster(x, ...)
recluster(x, ...)
# S3 method for class 'TDRObj'
recluster(
x,
.cl.resolution.parameter = 0.8,
.cl.method = "snn",
.small.size = 3,
.column.name = NULL,
.seed = 123,
.verbose = TRUE,
...
)Arguments
- x
A
TDRObj, Seurat, SingleCellExperiment, or HDF5AnnData (anndataR) object withget.graph()already run.- ...
Additional arguments passed to methods.
- .cl.resolution.parameter
Numeric resolution for Leiden CPM (default 0.8). Internally scaled by 1e-3. Higher = finer clusters.
- .cl.method
Character:
"snn"or"fgraph"(default"snn").- .small.size
Integer threshold for straggler absorption (default 3).
- .column.name
Character name for storing the solution (default
paste0("leiden.res.", .cl.resolution.parameter)). Cannot be"ids".- .seed
Integer for reproducibility (default 123).
- .verbose
Logical (default TRUE).