Subsampling function

This function subsamples a matrix using either random sampling or meta cells method.

subsampling(
  matrix,
  subsampling_method = c("sample", "meta_cells", "pseudobulk"),
  subsampling_ratio = 1,
  seed = 1,
  verbose = TRUE,
  ...
)

Arguments

matrix

The input matrix to be subsampled.

subsampling_method

The method to use for subsampling. Options are "sample", "pseudobulk" or "meta_cells".

subsampling_ratio

The percent of all samples used for fit_srm, default is 1.

seed

The random seed for cross-validation, default is 1.

verbose

Logical value, default is TRUE, whether to print progress messages.

...

Parameters for other methods.

Value

The subsampled matrix.

Examples

data("example_matrix")
data("example_ground_truth")
subsample_matrix <- subsampling(
  example_matrix,
  subsampling_ratio = 0.5
)
#> ℹ Subsample matrix generated, dimensions: 2500 cells by 18 genes.
subsample_matrix_2 <- subsampling(
  example_matrix,
  subsampling_method = "meta_cells",
  subsampling_ratio = 0.5,
  fast_pca = FALSE
)
#> ! number of PCs of PCA result is less than the desired number, using all PCs.
#> ℹ Subsample matrix generated, dimensions: 2500 cells by 18 genes.
subsample_matrix_3 <- subsampling(
  example_matrix,
  subsampling_method = "pseudobulk",
  subsampling_ratio = 0.5
)
#> ℹ Subsample matrix generated, dimensions: 2500 cells by 18 genes.

calculate_metrics(
  inferCSN(example_matrix),
  example_ground_truth,
  return_plot = TRUE
)
#> ℹ Running for <dense matrix>.
#> ℹ Checking input parameters.
#> ℹ Using L0 sparse regression model.
#> ℹ Using 1 core
#> ✔ Run done.
#> $metrics
#>      Metric  Value
#> 1     AUROC  0.952
#> 2     AUPRC  0.437
#> 3 Precision  0.529
#> 4    Recall  1.000
#> 5        F1  0.692
#> 6       ACC  0.948
#> 7        JI  0.514
#> 8        SI 18.000
#> 
#> $plot

#> 
calculate_metrics(
  inferCSN(subsample_matrix),
  example_ground_truth,
  return_plot = TRUE
)
#> ℹ Running for <dense matrix>.
#> ℹ Checking input parameters.
#> ℹ Using L0 sparse regression model.
#> ℹ Using 1 core
#> ✔ Run done.
#> $metrics
#>      Metric  Value
#> 1     AUROC  0.955
#> 2     AUPRC  0.449
#> 3 Precision  0.529
#> 4    Recall  1.000
#> 5        F1  0.692
#> 6       ACC  0.948
#> 7        JI  0.514
#> 8        SI 18.000
#> 
#> $plot

#> 
calculate_metrics(
  inferCSN(subsample_matrix_2),
  example_ground_truth,
  return_plot = TRUE
)
#> ℹ Running for <dense matrix>.
#> ℹ Checking input parameters.
#> ℹ Using L0 sparse regression model.
#> ℹ Using 1 core
#> ✔ Run done.
#> $metrics
#>      Metric  Value
#> 1     AUROC  0.952
#> 2     AUPRC  0.439
#> 3 Precision  0.529
#> 4    Recall  1.000
#> 5        F1  0.692
#> 6       ACC  0.948
#> 7        JI  0.514
#> 8        SI 18.000
#> 
#> $plot

#> 
calculate_metrics(
  inferCSN(subsample_matrix_3),
  example_ground_truth,
  return_plot = TRUE
)
#> ℹ Running for <dense matrix>.
#> ℹ Checking input parameters.
#> ℹ Using L0 sparse regression model.
#> ℹ Using 1 core
#> ✔ Run done.
#> $metrics
#>      Metric  Value
#> 1     AUROC  0.955
#> 2     AUPRC  0.449
#> 3 Precision  0.529
#> 4    Recall  1.000
#> 5        F1  0.692
#> 6       ACC  0.948
#> 7        JI  0.514
#> 8        SI 18.000
#> 
#> $plot

#>