Skip to contents

Calculate module activity scores

module_scores_sc.Rd

Implements the approach from Tirosh et al into bixverse, i.e., the AddModuleScore functionality from Seurat. For each module (gene set), computes the average expression of genes in the set minus the average expression of randomly selected control genes from the same expression bins. Genes are binned based on their average expression across cells to ensure controls are expression-matched.

Usage

module_scores_sc(
  object,
  gs_list,
  n_bins = 24L,
  n_ctrl = 100L,
  seed = 42L,
  streaming = NULL,
  .verbose = TRUE
)

Arguments

object

SingleCells class.

gs_list

Named list. The elements have the gene identifiers of the respective gene sets.

n_bins

Integer. Number of bins to use. Defaults to 24L.

n_ctrl

Integer. Number of control genes to use per gene for each gene set. Defaults to 100L.

seed

Integer. The random seed.

streaming

Optional Boolean. Shall the data be streamed in. Useful for larger data sets where you wish to avoid loading in the whole data. If NULL, will automatically detect.

.verbose

Boolean or integer. Controls verbosity and returns run times. FALSE -> quiet, TRUE or 1L -> normal verbosity, 2L -> detailed verbosity.

Value

Returns a ScMatrixRes with the module scores.

References

Tirosh et al, Science (2016)

Examples

# score the three planted marker programmes per cell
sc <- demo_single_cells()
gs_list <- list(
  type_1 = sprintf("gene_%02d", 1:10),
  type_2 = sprintf("gene_%02d", 11:20),
  type_3 = sprintf("gene_%02d", 21:30)
)
res <- module_scores_sc(sc, gs_list = gs_list, n_ctrl = 5L, .verbose = FALSE)
dim(res)
#> [1] 500   3

unlink(sc@dir_data, recursive = TRUE, force = TRUE)