Remove ambient RNA contamination from counts

This function provides a unified interface for ambient RNA correction using either SoupX or decontX. The input can be a Seurat object, a count matrix-like object, or a path that sn_read() can import.

Usage

sn_remove_ambient_contamination(
  x,
  raw = NULL,
  method = c("decontx", "soupx"),
  cluster = NULL,
  remove_zero_count_cells = FALSE,
  layer = "decontaminated_counts",
  return_object = TRUE,
  verbose = FALSE,
  ...
)

Arguments

x

A Seurat object, count matrix-like object, or path to filtered data.

raw

Optional raw/background counts. Required for method = "soupx". If supplied for decontx, it is used as the background matrix.

method

One of "decontx" or "soupx".

cluster

Optional cluster labels. This can be a vector with one value per cell, or a metadata column name when x is a Seurat object. If NULL, clusters are inferred with sn_run_cluster().

remove_zero_count_cells

Logical; if TRUE, remove cells whose decontX-corrected counts sum to zero. If FALSE, keep those cells by restoring their original counts and emit a warning.

layer

Layer name used when writing corrected counts back to a Seurat object. Defaults to "decontaminated_counts". Use "counts" to overwrite the original counts layer explicitly.

return_object

If TRUE and x is a Seurat object, return the updated Seurat object. Otherwise return the corrected counts matrix.

verbose

Logical; whether to print progress from helper clustering.

...

Additional method-specific arguments passed to celda::decontX() when method = "decontx", or to SoupX::autoEstCont() when method = "soupx".

Value

A corrected counts matrix, or an updated Seurat object when return_object = TRUE and x is a Seurat object. For decontX-based Seurat returns, the decontX_contamination and decontX_clusters columns are added to meta.data.

Examples

if (FALSE) { # \dontrun{
data("pbmc_small", package = "Shennong")
data("pbmc_small_raw", package = "Shennong")
pbmc <- sn_remove_ambient_contamination(pbmc, method = "decontx")

corrected <- sn_remove_ambient_contamination(
  x = SeuratObject::LayerData(pbmc, layer = "counts"),
  raw = pbmc_small_raw,
  method = "soupx",
  return_object = FALSE
)
} # }