Simulate gene expression data for biomarker panel testing

Simulates log-scale gene expression datasets with distributional shifts. Useful for benchmarking panel optimization routines under n << p conditions.

simulate_expression_data(
  p = 50L,
  n = 200L,
  k = 4L,
  seed = NULL,
  shift_scale = 0.6
)

Arguments

p

Number of features (genes).

n

Number of samples per dataset.

k

Number of datasets to simulate.

seed

Optional random seed for reproducibility.

shift_scale

Magnitude of dataset-specific feature shifts.

Value

A list with x_list (list of matrices), y_list (list of factors), and metadata (simulation parameters).