Sparse Partial Correlations On Gene Expression

---

Documentation for package ‘SPONGE’ version 1.30.0

• DESCRIPTION file.
• User guides, package vignettes and other documentation.
• Package NEWS.

Help Pages

build_classifier_central_genes	build classifiers for central genes
calibrate_model	tests and trains a model for a disease using a training and test data set (e.g., TCGA-BRCA and METABRIC)
ceRNA_interactions	ceRNA interactions
check_and_convert_expression_data	Checks if expression data is in matrix or ExpressionSet format and converts the latter to a standard matrix. Alternatively, a big.matrix descriptor object can be supplied to make use of shared memory between parallelized workers through the bigmemory package.
define_modules	Functions to define Sponge modules, created as all the first neighbors of the most central genes
enrichment_modules	Calculate enrichment scores
ensembl.df	example potential central nodes
filter_ceRNA_network	prepare ceRNA network and network centralities from SPONGE / SPONGEdb for spongEffects
fn_combined_centrality	Function to calculate centrality scores Calculation of combined centrality scores as proposed by Del Rio et al. (2009)
fn_discretize_spongeffects	discretize #' (functions taken from: Jerby-Arnon et al. 2018)
fn_elasticnet	Computes an elastic net model
fn_exact_match_summary	Calibrate classification method
fn_filter_network	Preprocessing ceRNA network
fn_gene_miRNA_F_test	Perform F test for gene-miRNA elastic net model
fn_get_model_coef	Extract the model coefficients from an elastic net model
fn_get_rss	Compute the residual sum of squares error for an elastic net model
fn_get_semi_random_OE	Function to calculate semi random enrichment scores of modules OE (functions taken from: Jerby-Arnon et al. 2018)
fn_get_shared_miRNAs	Identify miRNAs for which both genes have miRNA binding sites aka miRNA response elements in the competing endogeneous RNA hypothesis
fn_OE_module	Function to calculate enrichment scores of modules OE (functions taken from: Jerby-Arnon et al. 2018)
fn_RF_classifier	RF classification model
fn_weighted_degree	Function to calculate centrality scores Calculation of weighted degree scores based on Opsahl et al. (2010) Hyperparameter to tune: Alpha = 0 -> degree centrality as defined in Freeman, 1978 (number of edges).
genes_pairwise_combinations	Compute all pairwise interactions for a number of genes as indices
gene_expr	Gene expression test data set
get_central_modules	prepare ceRNA network and network centralities from SPONGE / SPONGEdb
mircode_ensg	mircode predicted miRNA gene interactions
mircode_symbol	mircode predicted miRNA gene interactions
mir_expr	miRNA expression test data set
mir_interactions	miRNA / gene interactions
plot_accuracy_sensitivity_specificity	list of plots for (1) accuracy and (2) sensitivity + specificity (see Boniolo and Hoffmann 2022 et al. Fig. 3a and Fig. 3b)
plot_confusion_matrices	plots the confusion matrix from spongEffects train_and_test() (see Boniolo and Hoffmann 2022 et al. Fig. 3a and Fig. 3b)
plot_density_scores	plots the density of the model scores for subtypes (see Boniolo and Hoffmann 2022 et al. Fig. 2)
plot_heatmaps	plots the heatmaps from training_and_test_model (see Boniolo and Hoffmann 2022 et al. Fig. 6)
plot_involved_miRNAs_to_modules	plots the heatmap of miRNAs invovled in the interactions of the modules (see Boniolo and Hoffmann 2022 et al. Fig. 7a)
plot_top_modules	plots the top x gini index modules (see Boniolo and Hoffmann 2022 et al. Figure 5)
precomputed_cov_matrices	covariance matrices under the null hypothesis that sensitivity correlation is zero
precomputed_null_model	A null model for testing purposes
prepare_metabric_for_spongEffects	prepare METABRIC formats for spongEffects
prepare_tcga_for_spongEffects	prepare TCGA formats for spongEffects
Random_spongEffects	build random classifiers
sample_zero_mscor_cov	Sampling zero multiple miRNA sensitivity covariance matrices
sample_zero_mscor_data	Sample mscor coefficients from pre-computed covariance matrices
sponge	Compute competing endogeneous RNA interactions using Sparse Partial correlations ON Gene Expression (SPONGE)
sponge_build_null_model	Build null model for p-value computation
sponge_compute_p_values	Compute p-values for SPONGE interactions
sponge_edge_centralities	Computes edge centralities
sponge_gene_miRNA_interaction_filter	Determine miRNA-gene interactions to be considered in SPONGE
sponge_network	Prepare a sponge network for plotting
sponge_node_centralities	Computes various node centralities
sponge_plot_network	Plot a sponge network
sponge_plot_network_centralities	plot node network centralities
sponge_plot_simulation_results	Plot simulation results for different null models
sponge_run_benchmark	run sponge benchmark where various settings, i.e. with or without regression, single or pooled miRNAs, are compared.
sponge_subsampling	Sponge subsampling
targetscan_ensg	targetscan predicted miRNA gene interactions
targetscan_symbol	targetscan predicted miRNA gene interactions
test_cancer_gene_expr	example test expression data for spongEffects
test_cancer_metadata	example test sample meta data for spongEffects
test_cancer_mir_expr	example test miRNA data for spongEffects
train_cancer_gene_expr	example training expression data for spongEffects
train_cancer_metadata	example training sample meta data for spongEffects
train_cancer_mir_expr	example training miRNA data for spongEffects
train_ceRNA_interactions	example train ceRNA interactions for spongEffects
train_network_centralities	example train network centralities for spongEffects