Run LIANA cell-cell communication analysis

Usage

RunLIANA(
  srt,
  group.by,
  method = c("natmi", "connectome", "logfc", "sca", "cellphonedb"),
  resource = NULL,
  assay = NULL,
  min_cells = 5,
  return_all = FALSE,
  backend = c("cpp", "r"),
  verbose = TRUE,
  species = c("human", "mouse"),
  consensus = c("auto", "rank", "aggregate", "none"),
  consensus_args = list(),
  ...
)

Arguments

srt: A Seurat object.
group.by: Metadata column defining cell groups. Passed to liana::liana_wrap() as idents_col.
method: LIANA methods to run. Defaults to LIANA's internal methods.
resource: LIANA ligand-receptor resource(s). If NULL, "Consensus" is used for human data and "MouseConsensus" for mouse data. This is a ligand-receptor resource choice, not a multi-method result consensus.
assay: Assay used by LIANA. If NULL, LIANA uses the default assay.
min_cells: Minimum cells per identity retained by LIANA.
return_all: Whether LIANA should return all possible interactions.
backend: Backend used only for result post-processing and unified CCC table aggregation. Upstream LIANA inference is unchanged.
verbose: Whether to print the message. Default is TRUE.
species: Species used to select the default ligand-receptor resource.
consensus: Multi-method result aggregation. "auto" uses rank_aggregate() when at least two methods are requested and does not invent a consensus for a single method. "rank" returns magnitude and specificity consensus ranks, "aggregate" uses liana_aggregate(), and "none" keeps only method-specific results.
consensus_args: Named arguments passed to the selected LIANA aggregation function.
...: Additional arguments passed to liana::liana_wrap().

Value

A Seurat object with LIANA results stored in srt@tools[["LIANA"]].