Converts expression settings to format in which the total number of potential ligands is reduced up to n top-predicted active ligands.

convert_expression_settings_evaluation Converts expression settings to format in which the total number of potential ligands is reduced up to n top-predicted active ligands.(useful for applications when a lot of ligands are potentially active, a lot of settings need to be predicted and a multi-ligand model is trained).

Usage

convert_settings_topn_ligand_prediction(setting, importances, model, n, normalization)

Arguments

setting

A list containing the following elements: .$name: name of the setting; .$from: name(s) of the ligand(s) active in the setting of interest; .$diffexp: data frame or tibble containing at least 3 variables= $gene, $lfc (log fold change treated vs untreated) and $qval (fdr-corrected p-value)

importances

A data frame containing at least folowing variables: $setting, $test_ligand, $ligand and one or more feature importance scores. $test_ligand denotes the name of a possibly active ligand, $ligand the name of the truely active ligand.

model

A model object of a classification object as e.g. generated via caret.

n

The top n number of ligands according to the ligand activity state prediction model will be considered as potential ligand for the generation of a new setting.

normalization

Way of normalization of the importance measures: "mean" (classifcal z-score) or "median" (modified z-score)

Value

A list with following elements: $name, $from, $response. $response will be a gene-named logical vector indicating whether the gene's transcription was influenced by the active ligand(s) in the setting of interest.

Examples

if (FALSE) { # \dontrun{
settings <- lapply(expression_settings_validation[1:5], convert_expression_settings_evaluation)
settings_ligand_pred <- convert_settings_ligand_prediction(settings, all_ligands = unlist(extract_ligands_from_settings(settings, combination = FALSE)), validation = TRUE, single = TRUE)

weighted_networks <- construct_weighted_networks(lr_network, sig_network, gr_network, source_weights_df)
ligands <- extract_ligands_from_settings(settings_ligand_pred, combination = FALSE)
ligand_target_matrix <- construct_ligand_target_matrix(weighted_networks, lr_network, ligands)
ligand_importances <- dplyr::bind_rows(lapply(settings_ligand_pred, get_single_ligand_importances, ligand_target_matrix))
evaluation <- evaluate_importances_ligand_prediction(ligand_importances, "median", "lda")

settings <- lapply(expression_settings_validation[5:10], convert_expression_settings_evaluation)
settings_ligand_pred <- convert_settings_ligand_prediction(settings, all_ligands = unlist(extract_ligands_from_settings(settings, combination = FALSE)), validation = FALSE, single = TRUE)
ligands <- extract_ligands_from_settings(settings_ligand_pred, combination = FALSE)
ligand_target_matrix <- construct_ligand_target_matrix(weighted_networks, lr_network, ligands)
ligand_importances <- dplyr::bind_rows(lapply(settings_ligand_pred, get_single_ligand_importances, ligand_target_matrix, known = FALSE))
settings <- lapply(settings, convert_settings_topn_ligand_prediction, importances = ligand_importances, model = evaluation$model, n = 3, normalization = "median")
} # }