Run metacell partitioning for single-cell data

Usage

RunMetaCell(
  srt,
  method = c("supercell", "seacells", "metacell"),
  assay = NULL,
  layer = "counts",
  gamma = 20,
  group.by = NULL,
  envname = NULL,
  conda = "auto",
  prefix = "Metacell",
  tool_name = "Metacell",
  verbose = TRUE,
  ...
)

Arguments

srt: A Seurat object.
method: Metacell construction method. One of "supercell", "seacells", or "metacell".
assay: Assay to use for metacell construction.
layer: Assay layer used to extract the count matrix.
gamma: Metacell granularity parameter. For SuperCell, larger values produce fewer metacells (typical range 10–50). For MetaCell, this is the K parameter controlling the number of metacells. For SEACells, the comparable parameter is passed via ... (e.g. n_metacells).
group.by: Optional metadata column used to build metacells within each group independently (e.g. by sample or cell type), preventing metacells from crossing group boundaries.
envname: Python environment name (SEACells only). Passed to reticulate::use_condaenv() when method = "seacells".
conda: Conda executable path (SEACells only).
prefix: Prefix for metadata columns written to srt.
tool_name: Name of the srt@tools entry.
verbose: Whether to print the message. Default is TRUE.
...: Additional arguments passed to the underlying metacell method.

Value

A metacell-level Seurat object. The original single-cell Seurat is stored in @misc[["original_srt"]] and the cell-to-metacell membership vector in @misc[["cell_membership"]]. The returned object can be passed directly to any scop function (standard_scop(), CellDimPlot(), etc.).

References

Baran, Y. et al. (2019). MetaCell: analysis of single-cell RNA-seq data using K-nn graph partitions. Genome Biology.

Bilous, M. et al. (2022). SuperCell: a versatile tool for single-cell data analysis. Genome Biology.

Persad, S. et al. (2023). SEACells infers transcriptional and epigenomic cellular states from single-cell genomics data. Nature Biotechnology.

Examples

data(pancreas_sub)
pancreas_sub <- standard_scop(
  pancreas_sub,
  nHVF = 500,
  linear_reduction_dims = 20,
  linear_reduction_dims_use = 1:20,
  nonlinear_reduction_dims = 2,
  verbose = FALSE
)

mc1 <- RunMetaCell(
  pancreas_sub,
  method = "supercell",
  gamma = 20
)

MetaCellPlot(mc1, group.by = "CellType")

mc2 <- RunMetaCell(
  pancreas_sub,
  method = "metacell",
  gamma = 20
)

MetaCellPlot(mc2, group.by = "CellType")