Simulate Single-Cell Data for Proportion Testing

Generate a synthetic Seurat object with sample-level replicates and condition-driven composition shifts, designed for DA testing workflows. Default settings generate richer cell-type diversity (including rare groups) and sample structure for more realistic proportion-test regression checks.

Usage

SimulateProportionData(
  n_genes = 2000L,
  samples_per_condition = 5L,
  cells_per_sample = 350L,
  conditions = c("Control", "Treatment"),
  cell_types = c(
    "Tcell",
    "Bcell",
    "Myeloid",
    "NK",
    "DC",
    "Mono",
    "Epithelial",
    "Fibro",
    "Mast"
  ),
  base_props = c(0.2, 0.14, 0.16, 0.1, 0.08, 0.1, 0.12, 0.07, 0.03),
  condition_prop_shift = c(0.07, -0.03, -0.03, -0.01, -0.01, -0.01, 0.01, 0,
    0.01),
  condition_prop_matrix = NULL,
  prop_concentration = 35,
  marker_genes_per_type = 30L,
  marker_fc = 4,
  condition_de_genes = 100L,
  condition_de_fc = 1.5,
  dispersion = 2,
  sample_prefix = "S",
  batch_levels = c("Batch1", "Batch2", "Batch3"),
  assay = "RNA",
  seed = 11,
  verbose = TRUE
)

Arguments

n_genes

Number of genes to simulate.

samples_per_condition

Number of biological samples per condition.

cells_per_sample

Number of cells per sample.

conditions

Condition labels.

cell_types

Cell type labels.

base_props

Baseline cell-type proportions for the first condition.

condition_prop_shift

Additive shift applied to baseline proportions for the second condition.

condition_prop_matrix

Optional condition-specific proportion matrix (rows=conditions, cols=cell types).

prop_concentration

Dirichlet concentration controlling sample-to-sample variability.

marker_genes_per_type

Number of marker genes for each cell type.

marker_fc

Fold-change multiplier for marker genes.

condition_de_genes

Number of global condition-affected genes.

condition_de_fc

Fold-change multiplier for global condition-affected genes.

dispersion

Negative binomial size parameter.

sample_prefix

Prefix for sample IDs.

batch_levels

Batch labels assigned to samples.

assay

Assay name for Seurat object creation.

seed

Random seed.

verbose

Whether to print messages.

Value

A Seurat object containing metadata columns: Sample, Condition, Batch, CellType, SubCellType, and Phase. Simulation truth is stored in srt@tools[["SimulateProportionData"]].

See also

RunProportionTest

Examples

if (FALSE) { # \dontrun{
srt_sim <- SimulateProportionData(
  n_genes = 1500,
  samples_per_condition = 4,
  cells_per_sample = 250,
  seed = 42
)

srt_sim <- RunProportionTest(
  srt_sim,
  group.by = "CellType",
  split.by = "Condition",
  sample.by = "Sample",
  proportion_method = "propeller",
  comparison = list(c("Control", "Treatment"))
)
} # }