Run doublet-calling with scDblFinder

Usage

RunscDblFinder(
  object,
  assay = "RNA",
  db_rate = ncol(srt)/1000 * 0.01,
  data_type = NULL,
  verbose = TRUE,
  srt = NULL,
  sample.by = NULL,
  seed = 1L,
  ...
)

Arguments

object: A Seurat object.
assay: Assay used for doublet calling.
db_rate: Expected doublet rate.
data_type: Optional CheckDataType result, used internally to avoid rescanning the count matrix.
verbose: Whether to print messages.
srt: Deprecated alias for object; supply exactly one of the two. It will be removed in scop 1.0.0.
sample.by: For ATAC, required metadata column identifying captures or libraries in which doublets can form. Biological conditions are not capture IDs. scDblFinder runs independently per capture (multiSampleMode = "split"). RNA behavior is unchanged.
seed: Random seed for the ATAC path, restoring the caller's RNG.
...: Arguments passed to other functions.

Value

A Seurat object with db.scDblFinder_score and db.scDblFinder_class updated only for selected-assay cells. For ATAC, Misc(assay, "scop_atac_doublets") records input cells, capture assignments, parameters, model output and dependency version; RNA-only cells are preserved.

Details

For ChromatinAssay input, use all raw peaks with feature aggregation (aggregateFeatures = TRUE, nfeatures = 25, processing = "normFeatures"). nfeatures and other model parameters can be supplied in .... With omitted db_rate, the ATAC path uses the per-capture expected rate; explicitly supply the appropriate rate for the capture chemistry. Scores/classes are predictions, not ground-truth doublet labels. No cells are removed. Existing upstream doublet labels can instead be used directly to select cells; running a second detector is not required.