Skip to contents

Single cell test data with a planted ambient profile

generate_cellsweep_test_data.Rd

This function generates synthetic data for CellSweep test purposes. Every real barcode is a two-component multinomial: a planted fraction alpha of its library comes from the soup, the rest from its own cell type profile. Empty droplets are pure soup at a much smaller library size, which is what the ambient profile is estimated off. Real barcodes come first, the empty droplets after.

Usage

generate_cellsweep_test_data(
  syn_data_params = params_sc_synthetic_cellsweep(),
  seed = 42L
)

Arguments

syn_data_params

List. Contains the parameters for the generation of the synthetic data, see: params_sc_synthetic_cellsweep().

seed

Integer. The seed for the generation of the synthetic data.

Value

List with the following items

  • counts - dgRMatrix with cells x genes.

  • obs - data.table with cell_id, cell_grp (NA for the empty droplets), sample_id, is_empty and alpha_true (NA for the empty droplets).

  • var - data.table that contains the gene information.

  • ambient_true - Numeric vector. The soup the empty droplets were drawn from, named by gene and summing to one.

  • celltype_profiles_true - Numeric matrix of cell types x genes. Each row sums to one.

Details

The empty droplets carry no cell type label, which is exactly what cellsweep_sc() keys off: barcodes that are neither empty nor annotated are excluded from the fit. Load the counts with a fully permissive params_sc_min_quality(), otherwise the ingest deletes the empty droplets the model trains on.

Examples

# a small synthetic experiment with a planted soup
data <- generate_cellsweep_test_data(
  syn_data_params = params_sc_synthetic_cellsweep(
    n_real = 60L,
    n_empty = 200L,
    n_genes = 60L
  )
)
dim(data$counts)
#> [1] 260  60
head(data$obs, 3)
#>     cell_id    cell_grp sample_id is_empty alpha_true
#>      <char>      <char>    <char>   <lgcl>      <num>
#> 1: cell_001 cell_type_1 sample_01    FALSE 0.37124013
#> 2: cell_002 cell_type_2 sample_01    FALSE 0.09390341
#> 3: cell_003 cell_type_3 sample_01    FALSE 0.23890524