Skip to contents

Constructor for SCENIC parameters

params_scenic.Rd

Constructor for SCENIC parameters

Usage

params_scenic(
  min_counts = 50L,
  min_cells = 0.03,
  learner_type = c("randomforest", "extratrees", "grnboost2"),
  gene_batch_strategy = c("correlated", "random"),
  gene_batch_size = NULL,
  n_pcs = 50L,
  n_subsample = 100000L,
  learner_params = list()
)

Arguments

min_counts

Integer. Minimum total counts a gene needs to be included in the analysis. Defaults to 50L.

min_cells

Numeric. Minimum proportion of cells (between 0 and 1) that must express a gene for it to be considered. Defaults to 0.03.

learner_type

String. Regression learner to use. One of c("randomforest", "extratrees", "grnboost2"). Defaults to "randomforest".

gene_batch_strategy

String. Strategy for grouping target genes into batches. Only used for "randomforest" and "extratrees" learners; ignored for "grnboost2". One of c("correlated", "random"). Defaults to "correlated".

gene_batch_size

Integer or NULL. Number of genes per batch. If NULL (default), the batch size is determined automatically. Ignored for "grnboost2". Defaults to NULL.

n_pcs

Integer. Number of PCs to use for the correlated gene batch strategy. Defaults to 50L.

n_subsample

Integer. Cell subsampling threshold for the correlated gene batch strategy. If the number of cells meets or exceeds this value, n_subsample cells are randomly selected prior to running randomised SVD. Defaults to 100000L.

learner_params

List. Optional overrides for the regression learner parameters. For "randomforest", see params_scenic_random_forest_defaults(). For "extratrees", see params_scenic_extra_trees_defaults(). For "grnboost2", see params_scenic_gradient_boosting_defaults(). Defaults to list().

Value

A named list with the following elements:

  • min_counts - Integer. Minimum total counts a gene needs to be included in the analysis. Defaults to 50L.

  • min_cells - Numeric. Minimum proportion of cells (between 0 and 1) that must express a gene for it to be considered. Defaults to 0.03.

  • learner_type - String. Regression learner to use. One of c("randomforest", "extratrees", "grnboost2"). Defaults to "randomforest".

  • gene_batch_strategy - String. Strategy for grouping target genes into batches. Only used for "randomforest" and "extratrees" learners; ignored for "grnboost2". One of c("correlated", "random"). Defaults to "correlated".

  • gene_batch_size - Integer or NULL. Number of genes per batch. If NULL (default), the batch size is determined automatically. Ignored for "grnboost2". Defaults to NULL.

  • n_pcs - Integer. Number of PCs to use for the correlated gene batch strategy. Defaults to 50L.

  • n_subsample - Integer. Cell subsampling threshold for the correlated gene batch strategy. If the number of cells meets or exceeds this value, n_subsample cells are randomly selected prior to running randomised SVD. Defaults to 100000L.

  • The elements of the base list, overridden by learner_params, spliced in at this position.