Train weights for the markers

Usage

train_weights(
  path_to_gs,
  exprs,
  level = 1,
  scaled = FALSE,
  clusters = NULL,
  range = c(1, 5),
  data_split = c(0.7, 0.2, 0.1),
  epochs = 20,
  batch_size = 32,
  run_weights_on_test = TRUE
)

Arguments

path_to_gs: Path to the gene set file without weights
exprs: The expression matrix, or a seurat object (rows: genes, columns: samples/cells)
level: The level of the gene sets to train weights for if you have multiple levels of gene sets.
scaled: Whether the expression matrix is scaled
clusters: A named vector of cluster ids If exprs is a seurat object, this is ignored. The cluster ids are taken from the seurat object.
range: The range of the weights
data_split: A vector of fractions for training, validation and testing. If only two fractions are provided, no testing set will be used.
epochs: The number of epochs to train
batch_size: The batch size
run_weights_on_test: Whether to run the weights on the test set. Requires that data_split has three elements.

Value

A data frame with the weights, that can be used directly by gs_prepare.