Compute Gene Isoform Entropy Matrix — sc_gene

Calculates normalized Shannon entropy for gene isoform expression across cells. Higher entropy indicates more diverse isoform usage, lower entropy indicates dominance by fewer isoforms.

Usage

sc_gene_entropy(
  sce,
  assay = "counts",
  gene_col = "gene_id",
  alpha = .Machine$double.xmin,
  min_counts_per_cell = 5,
  isoform_min_pct_cells = 0.05,
  isoform_cumulative_pct = 0.95,
  min_cell_fraction = 0.25,
  threads = 1,
  show_progress = interactive()
)

Arguments

sce: A SingleCellExperiment object
assay: Name of assay containing isoform counts (default: "counts")
gene_col: Column name in rowData containing gene identifiers (default: "gene_id")
alpha: Pseudocount added to avoid log(0) (default: .Machine$double.xmin)
min_counts_per_cell: Minimum total gene counts per cell to include (default: 5)
isoform_min_pct_cells: Minimum fraction of cells expressing each isoform (default: 0.05)
isoform_cumulative_pct: Keep top isoforms contributing to this cumulative proportion (default: 0.95)
min_cell_fraction: Minimum fraction of cells with valid entropy per gene (default: 0.25)
threads: Number of threads for parallel processing (default: 1)
show_progress: Logical indicating whether to show progress (default: TRUE if interactive)

Value

Matrix with genes as rows and cells as columns containing normalized entropy values (0-1).

Examples

sce <- scuttle::mockSCE(ncells = 50, ngenes = 30)
SummarizedExperiment::rowData(sce)$gene_id <- sort(
  paste0("gene", sample(1:9, nrow(sce), replace = TRUE))
)
res <- sc_gene_entropy(sce, threads = 2)