Deconvolute with cell2location — deconvolute_cell2location • spacedeconv

Runs the cell2location model (via Python/reticulate) and returns a deconvolution matrix, optionally rescaled to relative values.

Usage

deconvolute_cell2location(
  spatial_obj,
  signature = NULL,
  epochs = 30000,
  n_cell = 10,
  alpha = 20,
  gpu = TRUE,
  result_name = "cell2location",
  values = "relative"
)

Arguments

spatial_obj: SpatialExperiment.
signature: Signature/model input expected by cell2location.
epochs: Training epochs for the model.
n_cell: cell2location hyperparameter.
alpha: cell2location hyperparameter.
gpu: Logical; whether to use GPU for training.
result_name: Prefix used to label result columns (default: "cell2location").
values: "relative" to rescale to fractions, or "absolute" to keep raw outputs.