Get cell score by transferring gene weights from another slide By default, quantile normalization is used to ensure distribution match

Usage

getTransferCellScores(
  ref_obj,
  tar_obj,
  sigma_choice,
  use_quantile_normalization = TRUE,
  agg_cell_type = FALSE,
  gs_weight_threshold = 0,
  sigma_choice_tar = NULL,
  gene_score_type = c("PCA", "regression"),
  verbose = TRUE
)

Arguments

ref_obj: Reference object (where the gene weights will be obtained)
tar_obj: Target object (where the cell scores will be obtained)
sigma_choice: Sigma value to be used
use_quantile_normalization: Logical; apply quantile normalization of target to reference distribution (default TRUE)
agg_cell_type: Logical; if TRUE, returns a single matrix aggregated across cell types (default FALSE)
gs_weight_threshold: Numeric; absolute threshold used to zero out small gene weights in the gene score matrix prior to transfer (default 0).
sigma_choice_tar: Numeric; sigma value for target object. If NULL (default), uses sigma_choice. Not recommended for general use.
gene_score_type: Character; which gene score slot to use for transfer. "PCA" (default) uses the PCA back-projection weights in @geneScores. "regression" uses the regression-based weights in @geneScoresRegression, which avoids collinearity issues and produces more robust transfers. The regression slot must have been populated by calling computeRegressionGeneScores on the reference object.
verbose: verbose

Value

cell scores as a matrix