Find All Differentially Expressed Programs (1-vs-all analysis)

Uses scran::findMarkers to identify NMF programs that are differentially expressed in each cell type compared to all other cell types. Treats each NMF program as a "gene" and performs 1-vs-all comparisons.

Usage

FindAllDEPs(
  x,
  cell_type_col,
  nmf_name = "NMF",
  test = "wilcox",
  pval.type = "some",
  log.p = FALSE,
  min.prop = 0.1
)

Arguments

x

A SingleCellExperiment object with NMF results

cell_type_col

Character, column name in colData containing cell type labels

nmf_name

Character, name of NMF result to use (default "NMF")

test

Character, statistical test to use: "wilcox" (default), "t", "binom"

pval.type

Character, p-value type: "some", "any", "all" (default "some")

log.p

Logical, whether to log-transform p-values (default FALSE)

min.prop

Numeric, minimum proportion of cells that must express program (default 0.1)

Value

List of DataFrames (one per cell type) with differential expression statistics

Examples

# Run NMF and find differentially expressed programs
sce <- runNMFscape(sce, k = 8)
#> 
#> Attaching package: ‘Matrix’
#> The following object is masked from ‘package:S4Vectors’:
#> 
#>     expand
#> Error: object 'sce' not found
deps <- FindAllDEPs(sce, "cell_type")
#> Error: object 'sce' not found

# View results for first cell type
head(deps[[1]])
#> Error in h(simpleError(msg, call)): error in evaluating the argument 'x' in selecting a method for function 'head': object 'deps' not found