Package 'SPONGE' reference manual

Title:	Sparse Partial Correlations On Gene Expression
Description:	This package provides methods to efficiently detect competitive endogeneous RNA interactions between two genes. Such interactions are mediated by one or several miRNAs such that both gene and miRNA expression data for a larger number of samples is needed as input. The SPONGE package now also includes spongEffects: ceRNA modules offer patient-specific insights into the miRNA regulatory landscape.
Authors:	Markus List [aut, cre] , Markus Hoffmann [aut] , Lena Strasser [aut]
Maintainer:	Markus List <[email protected]>
License:	GPL (>=3)
Version:	1.29.0
Built:	2024-11-30 05:09:08 UTC
Source:	https://github.com/bioc/SPONGE

build classifiers for central genes

Description

build classifiers for central genes

Usage

build_classifier_central_genes(
  train_gene_expr,
  test_gene_expr,
  train_enrichment_modules,
  test_enrichment_modules,
  train_meta_data,
  test_meta_data,
  train_meta_data_type = "TCGA",
  test_meta_data_type = "TCGA",
  metric = "Exact_match",
  tunegrid_c = c(1:100),
  n.folds = 10,
  repetitions = 3
)
build_classifier_central_genes(
  train_gene_expr,
  test_gene_expr,
  train_enrichment_modules,
  test_enrichment_modules,
  train_meta_data,
  test_meta_data,
  train_meta_data_type = "TCGA",
  test_meta_data_type = "TCGA",
  metric = "Exact_match",
  tunegrid_c = c(1:100),
  n.folds = 10,
  repetitions = 3
)

Arguments

`train_gene_expr`	expression data of train dataset, genenames must be in rownames
`test_gene_expr`	expression data of test dataset, genenames must be in rownames
`train_enrichment_modules`	return of enrichment_modules()
`test_enrichment_modules`	return of enrichment_modules()
`train_meta_data`	meta data of train dataset
`test_meta_data`	meta data of test dataset
`train_meta_data_type`	TCGA or METABRIC
`test_meta_data_type`	TCGA or METABRIC
`metric`	metric (Exact_match, Accuracy) (default: Exact_match)
`tunegrid_c`	defines the grid for the hyperparameter optimization during cross validation (caret package) (default: 1:100)
`n.folds`	number of folds to be calculated
`repetitions`	number of k-fold cv iterations (default: 3)

Value

model for central genes

tests and trains a model for a disease using a training and test data set (e.g., TCGA-BRCA and METABRIC)

Description

tests and trains a model for a disease using a training and test data set (e.g., TCGA-BRCA and METABRIC)

Usage

calibrate_model(
  Input,
  modules_metadata,
  label,
  sampleIDs,
  Metric = "Exact_match",
  tunegrid_c = c(1:100),
  n_folds = 10,
  repetitions = 3
)
calibrate_model(
  Input,
  modules_metadata,
  label,
  sampleIDs,
  Metric = "Exact_match",
  tunegrid_c = c(1:100),
  n_folds = 10,
  repetitions = 3
)

Arguments

`Input`	Features to use for model calibration.
`modules_metadata`	metadata table containing information about samples/patients
`label`	Column of metadata to use as label in classification model
`sampleIDs`	Column of metadata containing sample/patient IDs to be matched with column names of spongEffects scores
`Metric`	metric (Exact_match, Accuracy) (default: Exact_match)
`tunegrid_c`	defines the grid for the hyperparameter optimization during cross validation (caret package) (default: 1:100)
`n_folds`	number of folds (default: 10)
`repetitions`	number of k-fold cv iterations (default: 3)
`modules`	return from enrichment_modules() function

Value

returns a list with the trained model and the prediction results Calibrate classification RF classification model

returns a list with the trained model and the prediction results

ceRNA interactions

Description

ceRNA interactions

Usage

ceRNA_interactions
ceRNA_interactions

Format

A data table of ceRNA interactions typically provided by sponge

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.

Description

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.

Usage

check_and_convert_expression_data(expr_data)
check_and_convert_expression_data(expr_data)

Arguments

expr_data

expr_data as matrix or ExpressionSet

Value

expr_data as matrix

Examples

## Not run: check_and_convert_expression_data(gene_expr)
## Not run: check_and_convert_expression_data(gene_expr)

Functions to define Sponge modules, created as all the first neighbors of the most central genes

Description

Functions to define Sponge modules, created as all the first neighbors of the most central genes

Usage

define_modules(
  network,
  central.modules = F,
  remove.central = T,
  set.parallel = T
)
define_modules(
  network,
  central.modules = F,
  remove.central = T,
  set.parallel = T
)

Arguments

`network`	Network as dataframe and list of central nodes. First two columns of the dataframe should contain the information of the nodes connected by edges.
`central.modules`	consider central gene as part of the module (default: False)
`remove.central`	Possibility of keeping or removing (default) central genes in the modules (default: T)
`set.parallel`	paralleling calculation of define_modules() (default: F)

Value

List of modules. Module names are the corresponding central genes.

Calculate enrichment scores

Description

Calculate enrichment scores

Usage

enrichment_modules(
  Expr.matrix,
  modules,
  bin.size = 100,
  min.size = 10,
  max.size = 200,
  min.expr = 10,
  method = "OE",
  cores = 1
)
enrichment_modules(
  Expr.matrix,
  modules,
  bin.size = 100,
  min.size = 10,
  max.size = 200,
  min.expr = 10,
  method = "OE",
  cores = 1
)

Arguments

`Expr.matrix`	ceRNA expression matrix
`modules`	Result of define_modules()
`bin.size`	bin size (default: 100)
`min.size`	minimum module size (default: 10)
`max.size`	maximum module size (default: 200)
`min.expr`	minimum expression (default: 10)
`method`	Enrichment to be used (Overall Enrichment: OE or Gene Set Variation Analysis: GSVA) (default: OE)
`cores`	number of cores to be used to calculate entichment scores with gsva or ssgsea methods. Default 1

Value

matrix containing module enrichment scores (module x samples)

example potential central nodes

Description

example potential central nodes

Usage

ensembl.df
ensembl.df

Format

(downloaded via biomaRt)

prepare ceRNA network and network centralities from SPONGE / SPONGEdb for spongEffects

Description

prepare ceRNA network and network centralities from SPONGE / SPONGEdb for spongEffects

Usage

filter_ceRNA_network(
  sponge_effects,
  Node_Centrality = NA,
  add_weighted_centrality = T,
  mscor.threshold = NA,
  padj.threshold = NA
)
filter_ceRNA_network(
  sponge_effects,
  Node_Centrality = NA,
  add_weighted_centrality = T,
  mscor.threshold = NA,
  padj.threshold = NA
)

Arguments

`sponge_effects`	the ceRNA network downloaded as R object from SPONGEdb (Hoffmann et al., 2021) or created by SPONGE (List et al., 2019) (ends with _sponge_results in the SPONGE vignette)
`Node_Centrality`	the network analysis downloaded as R object from SPONGEdb (Hoffmann et al., 2021) or created by SPONGE and containing centrality measures. (List et al., 2019) (ends with _networkAnalysis in the SPONGE vignette, you can also use your own network centrality measurements) if network_analysis is NA then the function only filters the ceRNA network, otherwise it will filter the given network centralities, but will not recalculate them based on the filtered ceRNA network.
`add_weighted_centrality`	calculate and add weighted centrality measures to previously available centralities. Default = T
`mscor.threshold`	mscor threshold to be filtered (default: NA)
`padj.threshold`	adjusted p-value to be filtered (default: NA)

Value

list of filtered ceRNA network and network centrailies. You can access it with list$objectname for further spongEffects steps

Function to calculate centrality scores Calculation of combined centrality scores as proposed by Del Rio et al. (2009)

Description

Function to calculate centrality scores Calculation of combined centrality scores as proposed by Del Rio et al. (2009)

Usage

fn_combined_centrality(CentralityMeasures)
fn_combined_centrality(CentralityMeasures)

Arguments

CentralityMeasures

dataframe with centrality score measures as columns and samples as rows

Value

Vector containing combined centrality scores

discretize #' (functions taken from: Jerby-Arnon et al. 2018)

Description

discretize #' (functions taken from: Jerby-Arnon et al. 2018)

Usage

fn_discretize_spongeffects(v, n.cat)
fn_discretize_spongeffects(v, n.cat)

Arguments

`v`	gene distance (defined by mother function OE module function)
`n.cat`	size of the bins (defined by mother function OE module function)

Value

discretized

Computes an elastic net model

Description

Computes an elastic net model

Usage

fn_elasticnet(x, y, alpha.step = 0.1)
fn_elasticnet(x, y, alpha.step = 0.1)

Arguments

`x`	miRNA expression matrix
`y`	gene expression vector
`alpha.step`	Step size for alpha, the tuning parameter for elastic net.

Value

The best model, i.e. the one for which the selected alpha yielded the smallest residual sum of squares error

Calibrate classification method

Description

Calibrate classification method

Usage

fn_exact_match_summary(data, lev = NULL, model = NULL)
fn_exact_match_summary(data, lev = NULL, model = NULL)

Arguments

`data`	Dataframe with module scores/covariates (modules x samples) AND outcome variable
`lev`	(default: NULL)
`model`	(default: NULL)

Value

Model and confusion matrix in a list

Preprocessing ceRNA network

Description

Preprocessing ceRNA network

Usage

fn_filter_network(network, mscor.threshold = 0.1, padj.threshold = 0.01)
fn_filter_network(network, mscor.threshold = 0.1, padj.threshold = 0.01)

Arguments

`network`	ceRNA network as data (typically present in the outputs of sponge)
`mscor.threshold`	mscor threshold (default 0.1)
`padj.threshold`	adjusted p-value threshold (default 0.01)

Value

filtered ceRNA network

Perform F test for gene-miRNA elastic net model

Description

Perform F test for gene-miRNA elastic net model

Usage

fn_gene_miRNA_F_test(g_expr, m_expr, model, p.adj.threshold = NULL)
fn_gene_miRNA_F_test(g_expr, m_expr, model, p.adj.threshold = NULL)

Arguments

`g_expr`	A gene expression matrix with samples in rows and genes in columns
`m_expr`	A miRNA expression matrix with samples in rows and genes in columns. Sample number and order has to agree with above gene expression matrix
`model`	A nested elastic net model to be tested
`p.adj.threshold`	Threshold for FDR corrected p-value

Value

return data frame with miRNA, fstat and adjusted p.value (BH).

Extract the model coefficients from an elastic net model

Description

Extract the model coefficients from an elastic net model

Usage

fn_get_model_coef(model)
fn_get_model_coef(model)

Arguments

model

An elastic net model

Value

A data frame with miRNAs and coefficients

Compute the residual sum of squares error for an elastic net model

Description

Compute the residual sum of squares error for an elastic net model

Usage

fn_get_rss(model, x, y)
fn_get_rss(model, x, y)

Arguments

`model`	The elastic net model
`x`	The miRNA expression
`y`	The gene expression

Value

the RSS

Function to calculate semi random enrichment scores of modules OE (functions taken from: Jerby-Arnon et al. 2018)

Description

Function to calculate semi random enrichment scores of modules OE (functions taken from: Jerby-Arnon et al. 2018)

Usage

fn_get_semi_random_OE(r, genes.dist.q, b.sign, num.rounds = 1000)
fn_get_semi_random_OE(r, genes.dist.q, b.sign, num.rounds = 1000)

Arguments

`r`	expression matrix
`genes.dist.q`	values of the genes after binning (result of binning)
`b.sign`	does the signature contain less than 2 genes? (controll parameter) (is set by mother function (OE module function))
`num.rounds`	number of rounds (default: 1000)

Value

random signature scores

Identify miRNAs for which both genes have miRNA binding sites aka miRNA response elements in the competing endogeneous RNA hypothesis

Description

Identify miRNAs for which both genes have miRNA binding sites aka miRNA response elements in the competing endogeneous RNA hypothesis

Usage

fn_get_shared_miRNAs(geneA, geneB, mir_interactions)
fn_get_shared_miRNAs(geneA, geneB, mir_interactions)

Arguments

`geneA`	The first gene
`geneB`	The second gene
`mir_interactions`	A named list of genes, where for each gene all miRNA interacting partners are listed

Value

A vector with shared RNAs of the two genes.

Function to calculate enrichment scores of modules OE (functions taken from: Jerby-Arnon et al. 2018)

Description

Function to calculate enrichment scores of modules OE (functions taken from: Jerby-Arnon et al. 2018)

Usage

fn_OE_module(
  NormCount,
  gene.sign,
  bin.size = 100,
  num.rounds = 1000,
  set_seed = 42
)
fn_OE_module(
  NormCount,
  gene.sign,
  bin.size = 100,
  num.rounds = 1000,
  set_seed = 42
)

Arguments

`NormCount`	normalized counts
`gene.sign`	significant genes
`bin.size`	bin size (default: 100)
`num.rounds`	number of rounds (default: 1000)
`set_seed`	seed size (default: 42)

Value

Signature scores

RF classification model

Description

RF classification model

Usage

fn_RF_classifier(
  Input.object,
  K,
  rep,
  metric = "Exact_match",
  tunegrid,
  set_seed = 42
)
fn_RF_classifier(
  Input.object,
  K,
  rep,
  metric = "Exact_match",
  tunegrid,
  set_seed = 42
)

Arguments

`Input.object`	data.frame made by predictors and dependent variable
`K`	number of folds (k-fold)
`rep`	number of times repeating the cross validation
`metric`	metric (Exact_match, Accuracy) (default: Exact_match)
`tunegrid`	defines the grid for the hyperparameter optimization during cross validation (caret package)
`set_seed`	set seed (default: 42)

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).

Description

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).

Usage

fn_weighted_degree(network, undirected = T, Alpha = 1)
fn_weighted_degree(network, undirected = T, Alpha = 1)

Arguments

`network`	Network formatted as a dataframe with three columns containing respectively node1, node2 and weights
`undirected`	directionality of the network (default: T)
`Alpha`	degree centrality as defined in Barrat et al., 2004 (default: 1)

Value

Dataframe containing information about nodes and their weighted centrality measure

Gene expression test data set

Description

Gene expression test data set

Usage

gene_expr
gene_expr

Format

A data frame of expression values with samples in columns and genes in rows

Compute all pairwise interactions for a number of genes as indices

Description

Compute all pairwise interactions for a number of genes as indices

Usage

genes_pairwise_combinations(number.of.genes)
genes_pairwise_combinations(number.of.genes)

Arguments

number.of.genes

Number of genes for which all pairwise interactions are needed

Value

data frame with one row per unique pairwise combination. To be used as input for the sponge method.

prepare ceRNA network and network centralities from SPONGE / SPONGEdb

Description

prepare ceRNA network and network centralities from SPONGE / SPONGEdb

Usage

get_central_modules(
  central_nodes,
  node_centrality,
  ceRNA_class = c("lncRNA", "circRNA", "protein_coding"),
  centrality_measure = "Weighted_Degree",
  cutoff = 1000
)
get_central_modules(
  central_nodes,
  node_centrality,
  ceRNA_class = c("lncRNA", "circRNA", "protein_coding"),
  centrality_measure = "Weighted_Degree",
  cutoff = 1000
)

Arguments

`central_nodes`	Vector containing Ensemble IDs of the chosen RNAs to use as central nodes for the modules.
`node_centrality`	output from filter_ceRNA_network() or own measurement, if own measurement taken, please provide node_centrality_column
`ceRNA_class`	default c("lncRNA","circRNA","protein_coding") (see http://www.ensembl.org/info/genome/genebuild/biotypes.html)
`centrality_measure`	Type of centrality measure to use. (Default: "Weighted_Degree", calculated in filter_ceRNA_network())
`cutoff`	the top cutoff modules will be returned (default: 1000)

Value

top cutoff modules, with selected RNAs as central genes

miRNA expression test data set

Description

miRNA expression test data set

Usage

mir_expr
mir_expr

Format

A data frame of expression values with samples in columns and miRNA in rows

miRNA / gene interactions

Description

miRNA / gene interactions

Usage

mir_interactions
mir_interactions

Format

A data frame of regression coefficients typically provided by sponge_gene_miRNA_interaction_filter

mircode predicted miRNA gene interactions

Description

mircode predicted miRNA gene interactions

Usage

mircode_ensg
mircode_ensg

Format

A matrix gene ensembl ids vs miRNA family names. >=1 if interaction is predicted, 0 otherwise

Source

http://www.mircode.org/download.php

mircode predicted miRNA gene interactions

Description

mircode predicted miRNA gene interactions

Usage

mircode_symbol
mircode_symbol

Format

A matrix gene symbols vs miRNA family names. >=1 if interaction is predicted, 0 otherwise

Source

http://www.mircode.org/download.php

list of plots for (1) accuracy and (2) sensitivity + specificity (see Boniolo and Hoffmann 2022 et al. Fig. 3a and Fig. 3b)

Description

list of plots for (1) accuracy and (2) sensitivity + specificity (see Boniolo and Hoffmann 2022 et al. Fig. 3a and Fig. 3b)

Usage

plot_accuracy_sensitivity_specificity(
  trained_model,
  central_genes_model = NA,
  all_expression_model = NA,
  random_model,
  training_dataset_name = "TCGA",
  testing_dataset_name = "TCGA",
  subtypes
)
plot_accuracy_sensitivity_specificity(
  trained_model,
  central_genes_model = NA,
  all_expression_model = NA,
  random_model,
  training_dataset_name = "TCGA",
  testing_dataset_name = "TCGA",
  subtypes
)

Arguments

`trained_model`	returned from train_and_test_model
`central_genes_model`	returned from build_classifier_central_genes()
`all_expression_model`	training and testing like central_genes_model but on ALL common expression data
`random_model`	returned from train_and_test_model using the randomization
`training_dataset_name`	name of training (e.g., TCGA)
`testing_dataset_name`	name of testing set (e.g., METABRIC)
`subtypes`	array of subtypes (e.g., c("Normal", "LumA", "LumB", "Her2", "Basal"))

Value

list of plots for (1) accuracy and (2) sensitivity + specificity

plots the confusion matrix from spongEffects train_and_test() (see Boniolo and Hoffmann 2022 et al. Fig. 3a and Fig. 3b)

Description

plots the confusion matrix from spongEffects train_and_test() (see Boniolo and Hoffmann 2022 et al. Fig. 3a and Fig. 3b)

Usage

plot_confusion_matrices(trained_model, subtypes.testing.factors)
plot_confusion_matrices(trained_model, subtypes.testing.factors)

Arguments

`trained_model`	returned from train_and_test_model
`subtypes_testing_factors`	subtypes of testing samples as factors

Value

plot of the confusion matrix

returns confusion matrix plots of the trained model

plots the density of the model scores for subtypes (see Boniolo and Hoffmann 2022 et al. Fig. 2)

Description

plots the density of the model scores for subtypes (see Boniolo and Hoffmann 2022 et al. Fig. 2)

Usage

plot_density_scores(trained_model, spongEffects, meta_data, label, sampleIDs)
plot_density_scores(trained_model, spongEffects, meta_data, label, sampleIDs)

Arguments

`trained_model`	returned from train_and_test_model
`spongEffects`	output of enrichment_modules()
`meta_data`	metadata of samples (retrieved from prepare_tcga_for_spongEffects() or from prepare_metabric_for_spongEffects())
`label`	Column of metadata to use as label in classification model
`sampleIDs`	Column of metadata containing sample/patient IDs to be matched with column names of spongEffects scores
`meta_data_type`	TCGA or METABRIC

Value

plots density scores for subtypes

plots the heatmaps from training_and_test_model (see Boniolo and Hoffmann 2022 et al. Fig. 6)

Description

plots the heatmaps from training_and_test_model (see Boniolo and Hoffmann 2022 et al. Fig. 6)

Usage

plot_heatmaps(
  trained_model,
  spongEffects,
  meta_data,
  label,
  sampleIDs,
  Modules_to_Plot = 2,
  show.rownames = F,
  show.colnames = F
)
plot_heatmaps(
  trained_model,
  spongEffects,
  meta_data,
  label,
  sampleIDs,
  Modules_to_Plot = 2,
  show.rownames = F,
  show.colnames = F
)

Arguments

`trained_model`	returned from train_and_test_model
`spongEffects`	output of enrichment_modules()
`meta_data`	metadata of samples (retrieved from prepare_tcga_for_spongEffects() or from prepare_metabric_for_spongEffects())
`label`	Column of metadata to use as label in classification model
`sampleIDs`	Column of metadata containing sample/patient IDs to be matched with column names of spongEffects scores
`Modules_to_Plot`	Number of modules to plot in the heatmap. Default = 2
`show.rownames`	Add row names (i.e. module names) to the heatmap. Default = F
`show.colnames`	Add column names (i.e. sample names) to the heatmap. Default = F

Value

ComplexHeatmap object NOT FUNCTIONAL

plots the heatmap of miRNAs invovled in the interactions of the modules (see Boniolo and Hoffmann 2022 et al. Fig. 7a)

Description

plots the heatmap of miRNAs invovled in the interactions of the modules (see Boniolo and Hoffmann 2022 et al. Fig. 7a)

Usage

plot_involved_miRNAs_to_modules(
  sponge_modules,
  trained_model,
  gene_mirna_candidates,
  k_modules = 25,
  filter_miRNAs = 3,
  bioMart_gene_symbol_columns = "hgnc_symbol",
  bioMart_gene_ensembl = "hsapiens_gene_ensembl",
  width = 5,
  length = 5,
  show_row_names = T,
  show_column_names = T,
  show_annotation_column = F,
  title = "Frequency",
  legend_height = 1.5,
  labels_gp_fontsize = 8,
  title_gp_fontsize = 8,
  legend_width = 3,
  column_title = "Module",
  row_title = "miRNA",
  row_title_gp_fontsize = 10,
  column_title_gp_fontsize = 10,
  row_names_gp_fontsize = 7,
  column_names_gp_fontsize = 7,
  column_names_rot = 45,
  unit = "cm"
)
plot_involved_miRNAs_to_modules(
  sponge_modules,
  trained_model,
  gene_mirna_candidates,
  k_modules = 25,
  filter_miRNAs = 3,
  bioMart_gene_symbol_columns = "hgnc_symbol",
  bioMart_gene_ensembl = "hsapiens_gene_ensembl",
  width = 5,
  length = 5,
  show_row_names = T,
  show_column_names = T,
  show_annotation_column = F,
  title = "Frequency",
  legend_height = 1.5,
  labels_gp_fontsize = 8,
  title_gp_fontsize = 8,
  legend_width = 3,
  column_title = "Module",
  row_title = "miRNA",
  row_title_gp_fontsize = 10,
  column_title_gp_fontsize = 10,
  row_names_gp_fontsize = 7,
  column_names_gp_fontsize = 7,
  column_names_rot = 45,
  unit = "cm"
)

Arguments

`sponge_modules`	result of define_modules()
`trained_model`	returned from train_and_test_model
`gene_mirna_candidates`	output of SPONGE or SPONGEdb (miRNAs_significance)
`k_modules`	top k modules to be shown (default: 25)
`filter_miRNAs`	min rowsum to be reach of miRNAs (default: 3.0)
`bioMart_gene_symbol_columns`	bioMart dataset column for gene symbols (e.g. human: hgnc_symbol, mouse: mgi_symbol) (default: hgnc_symbol)
`bioMart_gene_ensembl`	bioMart gene ensemble name (e.g., hsapiens_gene_ensembl)
`width`	the width of the heatmap (default: 5)
`length`	the length of the heatmap (default: 5)
`show_row_names`	show row names (default: T)
`show_column_names`	show column names (default: T)
`show_annotation_column`	add annotation column to columns (default: F)
`title`	the title of the plot (default: "Frequency")
`legend_height`	the height of the legend (default: 1.5)
`labels_gp_fontsize`	the font size of the labels (default: 8)
`title_gp_fontsize`	the font size of the title (default: 8)
`legend_width`	the width of the legend (default: 3)
`column_title`	the column title (default: "Module")
`row_title`	the title of the rows (default: "miRNA")
`row_title_gp_fontsize`	the font size of the row title (default: 10)
`column_title_gp_fontsize`	the font size of the column title (default: 10)
`row_names_gp_fontsize`	the font size of the row names (default: 7)
`column_names_gp_fontsize`	the font size of the column names (default: 7)
`column_names_rot`	the rotation angel of the column names (default: 45)
`unit`	either cm or inch (see ComplexHeatmap parameter)

Value

plot object

plots the top x gini index modules (see Boniolo and Hoffmann 2022 et al. Figure 5)

Description

plots the top x gini index modules (see Boniolo and Hoffmann 2022 et al. Figure 5)

Usage

plot_top_modules(
  trained_model,
  k_modules = 25,
  k_modules_red = 10,
  text_size = 16
)
plot_top_modules(
  trained_model,
  k_modules = 25,
  k_modules_red = 10,
  text_size = 16
)

Arguments

`trained_model`	returned from train_and_test_model
`k_modules`	top k modules to be shown (default: 25)
`k_modules_red`	top k modules shown in red - NOTE: must be smaller than k_modules (default: 10)
`text_size`	text size (default 16)
`bioMart_gene_symbol_columns`	bioMart dataset column for gene symbols (e.g. human: hgnc_symbol, mouse: mgi_symbol) (default: hgnc_symbol)
`bioMart_gene_ensembl`	bioMart gene ensemble name (e.g., hsapiens_gene_ensembl).

Value

plot object for lollipop plot

covariance matrices under the null hypothesis that sensitivity correlation is zero

Description

covariance matrices under the null hypothesis that sensitivity correlation is zero

Usage

precomputed_cov_matrices
precomputed_cov_matrices

Format

A list (different gene-gene correlations k) of lists (different number of miRNAs m) of covariance matrices

A null model for testing purposes

Description

A null model for testing purposes

Usage

precomputed_null_model
precomputed_null_model

Format

A list (different gene-gene correlations k) of lists (different number of miRNAs m) of sampled mscor values (100 each, computed from 100 samples)

prepare METABRIC formats for spongEffects

Description

prepare METABRIC formats for spongEffects

Usage

prepare_metabric_for_spongEffects(
  metabric_expression,
  metabric_metadata,
  subtypes_of_interest,
  bioMart_gene_ensembl = "hsapiens_gene_ensembl",
  bioMart_gene_symbol_columns = "hgnc_symbol"
)
prepare_metabric_for_spongEffects(
  metabric_expression,
  metabric_metadata,
  subtypes_of_interest,
  bioMart_gene_ensembl = "hsapiens_gene_ensembl",
  bioMart_gene_symbol_columns = "hgnc_symbol"
)

Arguments

`metabric_expression`	filepath to expression data in metabric format
`metabric_metadata`	filepath to metabric metadata in metabric format
`subtypes_of_interest`	array e.g., c("LumA", "LumB", "Her2", "Basal", "Normal")
`bioMart_gene_ensembl`	bioMart gene ensemble name (e.g., hsapiens_gene_ensembl). (See https://www.bioconductor.org/packages/release/bioc/vignettes/biomaRt/inst/doc/biomaRt.html) (default: hsapiens_gene_ensembl)
`bioMart_gene_symbol_columns`	bioMart dataset column for gene symbols (e.g. human: hgnc_symbol, mouse: mgi_symbol) (default: hgnc_symbol)

Value

list with metabric expression and metadata. You can access it with list$objectname for further spongEffects steps

prepare TCGA formats for spongEffects

Description

prepare TCGA formats for spongEffects

Usage

prepare_tcga_for_spongEffects(
  tcga_cancer_symbol,
  normal_ceRNA_expression_data,
  tumor_ceRNA_expression_data,
  normal_metadata,
  tumor_metadata,
  clinical_data,
  tumor_stages_of_interest,
  subtypes_of_interest
)
prepare_tcga_for_spongEffects(
  tcga_cancer_symbol,
  normal_ceRNA_expression_data,
  tumor_ceRNA_expression_data,
  normal_metadata,
  tumor_metadata,
  clinical_data,
  tumor_stages_of_interest,
  subtypes_of_interest
)

Arguments

`tcga_cancer_symbol`	e.g., BRCA for breast cancer
`normal_ceRNA_expression_data`	normal ceRNA expression data (same structure as input for SPONGE)
`tumor_ceRNA_expression_data`	tumor ceRNA expression data (same structure as input for SPONGE)
`normal_metadata`	metadata for normal samples (TCGA format style, needs to include column: sampleID, PATIENT_ID)
`tumor_metadata`	metadata for tumor samples (TCGA format style, needs to include column: sampleID, PATIENT_ID)
`clinical_data`	clinical data for all patients (TCGA format style, needs to include column: PATIENT_ID, AJCC_PATHOLOGIC_TUMOR_STAGE)
`tumor_stages_of_interest`	array e.g., c(STAGE I', 'STAGE IA', 'STAGE IB', 'STAGE II', 'STAGE IIA')
`subtypes_of_interest`	array e.g., c("LumA", "LumB", "Her2", "Basal", "Normal")

Value

list of prepared data. You can access it with list$objectname for further spongEffects steps

build random classifiers

Description

build random classifiers

Usage

Random_spongEffects(
  sponge_modules,
  gene_expr,
  min.size = 10,
  bin.size = 100,
  max.size = 200,
  min.expression = 10,
  replace = F,
  method = "OE",
  cores = 1
)
Random_spongEffects(
  sponge_modules,
  gene_expr,
  min.size = 10,
  bin.size = 100,
  max.size = 200,
  min.expression = 10,
  replace = F,
  method = "OE",
  cores = 1
)

Arguments

`sponge_modules`	result of define_modules()
`gene_expr`	Input expression matri
`min.size`	minimum module size (default: 10)
`bin.size`	bin size (default: 100)
`max.size`	maximum module size (default: 200)
`replace`	Possibility of keeping or removing (default) central genes in the modules (default: F)
`method`	Enrichment to be used (Overall Enrichment: OE or Gene Set Variation Analysis: GSVA) (default: OE)
`cores`	number of cores to be used to calculate entichment scores with gsva or ssgsea methods. Default 1
`train_gene_expr`	expression data of train dataset, genenames must be in rownames
`test_gene_expr`	expression data of test dataset, genenames must be in rownames
`train_meta_data`	meta data of train dataset
`test_meta_data`	meta data of test dataset
`train_meta_data_type`	TCGA or METABRIC
`test_meta_data_type`	TCGA or METABRIC
`metric`	metric (Exact_match, Accuracy) (default: Exact_match)
`tunegrid_c`	defines the grid for the hyperparameter optimization during cross validation (caret package) (default: 1:100)
`n.folds`	number of folds to be calculated
`repetitions`	number of k-fold cv iterations (default: 3)
`min.expr`	minimum expression (default: 10)

Value

randomized prediction model Define random modules

A list with randomly defined modules and related enrichment scores

Sampling zero multiple miRNA sensitivity covariance matrices

Description

Sampling zero multiple miRNA sensitivity covariance matrices

Usage

sample_zero_mscor_cov(
  m,
  number_of_solutions,
  number_of_attempts = 1000,
  gene_gene_correlation = NULL,
  random_seed = NULL,
  log.level = "ERROR"
)
sample_zero_mscor_cov(
  m,
  number_of_solutions,
  number_of_attempts = 1000,
  gene_gene_correlation = NULL,
  random_seed = NULL,
  log.level = "ERROR"
)

Arguments

`m`	number of miRNAs, i.e. number of columns of the matrix
`number_of_solutions`	stop after this many instances have been samples
`number_of_attempts`	give up after that many attempts
`gene_gene_correlation`	optional, define the correlation of the first two elements, i.e. the genes.
`random_seed`	A random seed to be used for reproducible results
`log.level`	the log level, typically set to INFO, set to DEBUG for verbose logging

Value

a list of covariance matrices with zero sensitivity correlation

Examples

sample_zero_mscor_cov(m = 1,
number_of_solutions = 1,
gene_gene_correlation = 0.5)
sample_zero_mscor_cov(m = 1,
number_of_solutions = 1,
gene_gene_correlation = 0.5)

Sample mscor coefficients from pre-computed covariance matrices

Description

Sample mscor coefficients from pre-computed covariance matrices

Usage

sample_zero_mscor_data(
  cov_matrices,
  number_of_samples = 100,
  number_of_datasets = 100
)
sample_zero_mscor_data(
  cov_matrices,
  number_of_samples = 100,
  number_of_datasets = 100
)

Arguments

`cov_matrices`	a list of pre-computed covariance matrices
`number_of_samples`	the number of samples available in the expression data
`number_of_datasets`	the number of mscor coefficients to be sampled from each covariance matrix

Value

a vector of mscor coefficients

Examples

#we select from the pre-computed covariance matrices in SPONGE
#100 for m = 5 miRNAs and gene-gene correlation 0.6
cov_matrices_selected <- precomputed_cov_matrices[["5"]][["0.6"]]
sample_zero_mscor_data(cov_matrices = cov_matrices_selected,
number_of_samples = 200, number_of_datasets = 10)
#we select from the pre-computed covariance matrices in SPONGE
#100 for m = 5 miRNAs and gene-gene correlation 0.6
cov_matrices_selected <- precomputed_cov_matrices[["5"]][["0.6"]]
sample_zero_mscor_data(cov_matrices = cov_matrices_selected,
number_of_samples = 200, number_of_datasets = 10)

Compute competing endogeneous RNA interactions using Sparse Partial correlations ON Gene Expression (SPONGE)

Description

Compute competing endogeneous RNA interactions using Sparse Partial correlations ON Gene Expression (SPONGE)

Usage

sponge(
  gene_expr,
  mir_expr,
  mir_interactions = NULL,
  log.level = "ERROR",
  log.every.n = 1e+05,
  log.file = NULL,
  selected.genes = NULL,
  gene.combinations = NULL,
  each.miRNA = FALSE,
  min.cor = 0.1,
  parallel.chunks = 1000,
  random_seed = NULL,
  result_as_dt = FALSE
)
sponge(
  gene_expr,
  mir_expr,
  mir_interactions = NULL,
  log.level = "ERROR",
  log.every.n = 1e+05,
  log.file = NULL,
  selected.genes = NULL,
  gene.combinations = NULL,
  each.miRNA = FALSE,
  min.cor = 0.1,
  parallel.chunks = 1000,
  random_seed = NULL,
  result_as_dt = FALSE
)

Arguments

`gene_expr`	A gene expression matrix with samples in rows and featurs in columns. Alternatively an object of class ExpressionSet.
`mir_expr`	A miRNA expression matrix with samples in rows and features in columns. Alternatively an object of class ExpressionSet.
`mir_interactions`	A named list of genes, where for each gene we list all miRNA interaction partners that should be considered.
`log.level`	The log level, can be one of "info", "debug", "error"
`log.every.n`	write to the log after every n steps
`log.file`	write log to a file, particularly useful for paralleliyzation
`selected.genes`	Operate only on a subset of genes, particularly useful for bootstrapping
`gene.combinations`	A data frame of combinations of genes to be tested. Gene names are taken from the first two columns and have to match the names used for gene_expr
`each.miRNA`	Whether to consider individual miRNAs or pooling them.
`min.cor`	Consider only gene pairs with a minimum correlation specified here.
`parallel.chunks`	Split into this number of tasks if parallel processing is set up. The number should be high enough to guarantee equal distribution of the work load in parallel execution. However, if the number is too large, e.g. in the worst case one chunk per computation, the overhead causes more computing time than can be saved by parallel execution. Register a parallel backend that is compatible with foreach to use this feature. More information can be found in the documentation of the foreach / doParallel packages.
`random_seed`	A random seed to be used for reproducible results
`result_as_dt`	whether to return results as data table or data frame

Value

A data frame with significant gene-gene competetive endogenous RNA or 'sponge' interactions

Examples

#First, extract miRNA candidates for each of the genes
#using sponge_gene_miRNA_interaction_filter. Here we use a prepared
#dataset mir_interactions.

#Second we compute ceRNA interactions for all pairwise combinations of genes
#using all miRNAs remaining after filtering through elasticnet.
ceRNA_interactions <- sponge(
gene_expr = gene_expr,
mir_expr = mir_expr,
mir_interactions = mir_interactions)
#First, extract miRNA candidates for each of the genes
#using sponge_gene_miRNA_interaction_filter. Here we use a prepared
#dataset mir_interactions.

#Second we compute ceRNA interactions for all pairwise combinations of genes
#using all miRNAs remaining after filtering through elasticnet.
ceRNA_interactions <- sponge(
gene_expr = gene_expr,
mir_expr = mir_expr,
mir_interactions = mir_interactions)

Build null model for p-value computation

Description

Build null model for p-value computation

Usage

sponge_build_null_model(
  number_of_datasets = 1e+05,
  number_of_samples,
  cov_matrices = precomputed_cov_matrices,
  ks = seq(0.2, 0.9, 0.1),
  m_max = 8,
  log.level = "ERROR"
)
sponge_build_null_model(
  number_of_datasets = 1e+05,
  number_of_samples,
  cov_matrices = precomputed_cov_matrices,
  ks = seq(0.2, 0.9, 0.1),
  m_max = 8,
  log.level = "ERROR"
)

Arguments

`number_of_datasets`	the number of datesets defining the precision of the p-value
`number_of_samples`	the number of samples in the expression data
`cov_matrices`	pre-computed covariance matrices
`ks`	a sequence of gene-gene correlation values for which null models are computed
`m_max`	null models are build for each elt in ks for 1 to m_max miRNAs
`log.level`	The log level of the logging package

Value

a list (for various values of m) of lists (for various values of k) of lists of simulated data sets, drawn from a set of precomputed covariance matrices

Examples

sponge_build_null_model(100, 100,
cov_matrices = precomputed_cov_matrices[1:3], m_max = 3)
sponge_build_null_model(100, 100,
cov_matrices = precomputed_cov_matrices[1:3], m_max = 3)

Compute p-values for SPONGE interactions

Description

This method uses pre-computed covariance matrices that were created for various gene-gene correlations (0.2 to 0.9 in steps of 0.1) and number of miRNAs (between 1 and 8) under the null hypothesis that the sensitivity correlation is zero. Datasets are sampled from this null model and allow for an empirical p-value to be computed that is only significant if the sensitivity correlation is higher than can be expected by chance given the number of samples, correlation and number of miRNAs. p-values are adjusted indepdenently for each parameter combination using Benjamini-Hochberg FDR correction.

Usage

sponge_compute_p_values(sponge_result, null_model, log.level = "ERROR")
sponge_compute_p_values(sponge_result, null_model, log.level = "ERROR")

Arguments

`sponge_result`	A data frame from a sponge call
`null_model`	optional, pre-computed simulated data
`log.level`	The log level of the logging package

Value

A data frame with sponge results, now including p-values and adjusted p-value

Examples

sponge_compute_p_values(ceRNA_interactions,
null_model = precomputed_null_model)
sponge_compute_p_values(ceRNA_interactions,
null_model = precomputed_null_model)

Computes edge centralities

Description

Computes edge betweenness centrality for the ceRNA interaction network induced by the results of the SPONGE method.

Usage

sponge_edge_centralities(sponge_result)
sponge_edge_centralities(sponge_result)

Arguments

sponge_result

The output generated by the sponge method.

Value

data table or data frame with gene, degree, eigenvector and betweenness

Examples

sponge_edge_centralities(ceRNA_interactions)
sponge_edge_centralities(ceRNA_interactions)

Determine miRNA-gene interactions to be considered in SPONGE

Description

The purpose of this method is to limit the number of miRNA-gene interactions we need to consider in SPONGE. There are 3 filtering steps: 1. variance filter (optional). Only considre genes and miRNAs with variance > var.threshold. 2. miRNA target database filter (optional). Use a miRNA target database provided by the user to filter for those miRNA gene interactions for which evidence exists. This can either be predicted target interactions or experimentally validated ones. 3. For each remaining interaction of a gene and its regulating miRNAs use elastic net regression to achieve a) Feature selection: We only retain miRNAs that influence gene expression b) Effect strength: The sign of the coefficients allows us to filter for miRNAs that down-regulate gene expression. Moreover, we can use the coefficients to rank the miRNAs by their relative effect strength. We strongly recommend setting up a parallel backend compatible with the foreach package. See example and the documentation of the foreach and doParallel packages.

Usage

sponge_gene_miRNA_interaction_filter(
  gene_expr,
  mir_expr,
  mir_predicted_targets,
  elastic.net = TRUE,
  log.level = "ERROR",
  log.file = NULL,
  var.threshold = NULL,
  F.test = FALSE,
  F.test.p.adj.threshold = 0.05,
  coefficient.threshold = -0.05,
  coefficient.direction = "<",
  select.non.targets = FALSE,
  random_seed = NULL,
  parallel.chunks = 100
)
sponge_gene_miRNA_interaction_filter(
  gene_expr,
  mir_expr,
  mir_predicted_targets,
  elastic.net = TRUE,
  log.level = "ERROR",
  log.file = NULL,
  var.threshold = NULL,
  F.test = FALSE,
  F.test.p.adj.threshold = 0.05,
  coefficient.threshold = -0.05,
  coefficient.direction = "<",
  select.non.targets = FALSE,
  random_seed = NULL,
  parallel.chunks = 100
)

Arguments

`gene_expr`	A gene expression matrix with samples in rows and featurs in columns. Alternatively an object of class ExpressionSet.
`mir_expr`	A miRNA expression matrix with samples in rows and features in columns. Alternatively an object of class ExpressionSet.
`mir_predicted_targets`	A data frame with miRNA in cols and genes in rows. A 0 indicates the miRNA is not predicted to target the gene, >0 otherwise. If this parameter is NULL all miRNA-gene interactions are tested
`elastic.net`	Whether to apply elastic net regression filtering or not.
`log.level`	One of 'warn', 'error', 'info'
`log.file`	Log file to write to
`var.threshold`	Only consider genes and miRNA with variance > var.threshold. If this parameter is NULL no variance filtering is performed.
`F.test`	If true, an F-test is performed on each model parameter to assess its importance for the model based on the RSS of the full model vs the RSS of the nested model without the miRNA in question. This is time consuming and has the potential disadvantage that correlated miRNAs are removed even though they might play a role in ceRNA interactions. Use at your own risk.
`F.test.p.adj.threshold`	If F.test is TRUE, threshold to use for miRNAs to be included.
`coefficient.threshold`	threshold to cross for a regression coefficient to be called significant. depends on the parameter coefficient.direction.
`coefficient.direction`	If "<", coefficient has to be lower than coefficient.threshold, if ">", coefficient has to be larger than threshold. If NULL, the absolute value of the coefficient has to be larger than the threshold.
`select.non.targets`	For testing effect of miRNA target information. If TRUE, the method determines as usual which miRNAs are potentially targeting a gene. However, these are then replaced by a random sample of non-targeting miRNAs (without seeds) of the same size. Useful for testing if observed effects are caused by miRNA regulation.
`random_seed`	A random seed to be used for reproducible results
`parallel.chunks`	Split into this number of tasks if parallel processing is set up. The number should be high enough to guarantee equal distribution of the work load in parallel execution. However, if the number is too large, e.g. in the worst case one chunk per computation, the overhead causes more computing time than can be saved by parallel execution. Register a parallel backend that is compatible with foreach to use this feature. More information can be found in the documentation of the foreach / doParallel packages.

Value

A list of genes, where for each gene, the regulating miRNA are included as a data frame. For F.test = TRUE this is a data frame with fstat and p-value for each miRNA. Else it is a data frame with the model coefficients.

Examples

#library(doParallel)
#cl <- makePSOCKcluster(2)
#registerDoParallel(cl)
genes_miRNA_candidates <- sponge_gene_miRNA_interaction_filter(
gene_expr = gene_expr,
mir_expr = mir_expr,
mir_predicted_targets = targetscan_symbol)
#stopCluster(cl)

#If we also perform an F-test, only few of the above miRNAs remain
genes_miRNA_candidates <- sponge_gene_miRNA_interaction_filter(
gene_expr = gene_expr,
mir_expr = mir_expr,
mir_predicted_targets = targetscan_symbol,
F.test = TRUE,
F.test.p.adj.threshold = 0.05)

#library(doParallel)
#cl <- makePSOCKcluster(2)
#registerDoParallel(cl)
genes_miRNA_candidates <- sponge_gene_miRNA_interaction_filter(
gene_expr = gene_expr,
mir_expr = mir_expr,
mir_predicted_targets = targetscan_symbol)
#stopCluster(cl)

#If we also perform an F-test, only few of the above miRNAs remain
genes_miRNA_candidates <- sponge_gene_miRNA_interaction_filter(
gene_expr = gene_expr,
mir_expr = mir_expr,
mir_predicted_targets = targetscan_symbol,
F.test = TRUE,
F.test.p.adj.threshold = 0.05)

Prepare a sponge network for plotting

Description

Prepare a sponge network for plotting

Usage

sponge_network(
  sponge_result,
  mir_data,
  target.genes = NULL,
  show.sponge.interaction = TRUE,
  show.mirnas = "none",
  min.interactions = 3
)
sponge_network(
  sponge_result,
  mir_data,
  target.genes = NULL,
  show.sponge.interaction = TRUE,
  show.mirnas = "none",
  min.interactions = 3
)

Arguments

`sponge_result`	ceRNA interactions as produced by the sponge method.
`mir_data`	miRNA interactions as produced by sponge_gene_miRNA_interaction_filter
`target.genes`	a character vector to select a subset of genes
`show.sponge.interaction`	whether to connect ceRNAs
`show.mirnas`	one of none, shared, all
`min.interactions`	minimum degree of a gene to be shown

Value

a list of nodes and edges

Examples

sponge_network(ceRNA_interactions, mir_interactions)
sponge_network(ceRNA_interactions, mir_interactions)

Computes various node centralities

Description

Computes degree, eigenvector centrality and betweenness centrality for the ceRNA interaction network induced by the results of the SPONGE method

Usage

sponge_node_centralities(sponge_result, directed = FALSE)
sponge_node_centralities(sponge_result, directed = FALSE)

Arguments

`sponge_result`	output of the sponge method
`directed`	Whether to consider the input network as directed or not.

Value

data table or data frame with gene, degree, eigenvector and betweenness

Examples

sponge_node_centralities(ceRNA_interactions)
sponge_node_centralities(ceRNA_interactions)

Plot a sponge network

Description

Plot a sponge network

Usage

sponge_plot_network(
  sponge_result,
  mir_data,
  layout = "layout.fruchterman.reingold",
  force.directed = FALSE,
  ...
)
sponge_plot_network(
  sponge_result,
  mir_data,
  layout = "layout.fruchterman.reingold",
  force.directed = FALSE,
  ...
)

Arguments

`sponge_result`	ceRNA interactions as produced by the sponge method.
`mir_data`	miRNA interactions as produced by sponge_gene_miRNA_interaction_filter
`layout`	one of the layout methods supported in the visNetwork package
`force.directed`	whether to produce a force directed network, gets slow for large networks
`...`	further params for sponge_network

Value

shows a plot

Examples

sponge_plot_network(ceRNA_interactions, mir_interactions)
sponge_plot_network(ceRNA_interactions, mir_interactions)

plot node network centralities

Description

plot node network centralities

Usage

sponge_plot_network_centralities(
  network_centralities,
  measure = "all",
  x = "degree",
  top = 5,
  base_size = 18
)
sponge_plot_network_centralities(
  network_centralities,
  measure = "all",
  x = "degree",
  top = 5,
  base_size = 18
)

Arguments

`network_centralities`	a result from sponge_node_centralities()
`measure`	one of 'all', 'degree', 'ev' or 'btw'
`x`	plot against another column in the data table, defaults to degree
`top`	label the top x samples in the plot
`base_size`	size of the text in the plot

Value

a plot

Examples

## Not run: 
network_centralities <- sponge_node_centralities(ceRNA_interactions)
sponge_plot_network_centralities(network_centralities)
## End(Not run)
## Not run: 
network_centralities <- sponge_node_centralities(ceRNA_interactions)
sponge_plot_network_centralities(network_centralities)
## End(Not run)

Plot simulation results for different null models

Description

Plot simulation results for different null models

Usage

sponge_plot_simulation_results(null_model_data)
sponge_plot_simulation_results(null_model_data)

Arguments

null_model_data

the output of sponge_build_null_model

Value

a ggplot2 object

Examples

sponge_plot_simulation_results(precomputed_null_model)
sponge_plot_simulation_results(precomputed_null_model)

run sponge benchmark where various settings, i.e. with or without regression, single or pooled miRNAs, are compared.

Description

run sponge benchmark where various settings, i.e. with or without regression, single or pooled miRNAs, are compared.

Usage

sponge_run_benchmark(
  gene_expr,
  mir_expr,
  mir_predicted_targets,
  number_of_samples = 100,
  number_of_datasets = 100,
  number_of_genes_to_test = c(25),
  compute_significance = FALSE,
  folder = NULL
)
sponge_run_benchmark(
  gene_expr,
  mir_expr,
  mir_predicted_targets,
  number_of_samples = 100,
  number_of_datasets = 100,
  number_of_genes_to_test = c(25),
  compute_significance = FALSE,
  folder = NULL
)

Arguments

`gene_expr`	A gene expression matrix with samples in rows and featurs in columns. Alternatively an object of class ExpressionSet.
`mir_expr`	A miRNA expression matrix with samples in rows and features in columns. Alternatively an object of class ExpressionSet.
`mir_predicted_targets`	(a list of) mir interaction sources such as targetscan, etc.
`number_of_samples`	number of samples in the null model
`number_of_datasets`	number of datasets to sample from the null model
`number_of_genes_to_test`	a vector of numbers of genes to be tested, e.g. c(250,500)
`compute_significance`	whether to compute p-values
`folder`	where the results should be saved, if NULL no output to disk

Value

a list (regression, no regression) of lists (single miRNA, pooled miRNAs) of benchmark results

Examples

sponge_run_benchmark(gene_expr = gene_expr, mir_expr = mir_expr,
mir_predicted_targets = targetscan_symbol,
number_of_genes_to_test = c(10), folder = NULL)
sponge_run_benchmark(gene_expr = gene_expr, mir_expr = mir_expr,
mir_predicted_targets = targetscan_symbol,
number_of_genes_to_test = c(10), folder = NULL)

Sponge subsampling

Description

Sponge subsampling

Usage

sponge_subsampling(
  subsample.n = 100,
  subsample.repeats = 10,
  subsample.with.replacement = FALSE,
  subsample.plot = FALSE,
  gene_expr,
  mir_expr,
  ...
)
sponge_subsampling(
  subsample.n = 100,
  subsample.repeats = 10,
  subsample.with.replacement = FALSE,
  subsample.plot = FALSE,
  gene_expr,
  mir_expr,
  ...
)

Arguments

`subsample.n`	the number of samples to be drawn in each round
`subsample.repeats`	how often should the subsampling be done?
`subsample.with.replacement`	logical, should we allow samples to be used repeatedly
`subsample.plot`	logical, should the results be plotted as box plots
`gene_expr`	A gene expression matrix with samples in rows and featurs in columns. Alternatively an object of class ExpressionSet.
`mir_expr`	A miRNA expression matrix with samples in rows and features in columns. Alternatively an object of class ExpressionSet.
`...`	parameters passed on to the sponge function

Value

a summary of the results with mean and standard deviations of the correlation and sensitive correlation.

References

sponge

Examples

sponge_subsampling(gene_expr = gene_expr,
mir_expr = mir_expr, mir_interactions = mir_interactions,
subsample.n = 10, subsample.repeats = 1)
sponge_subsampling(gene_expr = gene_expr,
mir_expr = mir_expr, mir_interactions = mir_interactions,
subsample.n = 10, subsample.repeats = 1)

targetscan predicted miRNA gene interactions

Description

targetscan predicted miRNA gene interactions

Usage

targetscan_ensg
targetscan_ensg

Format

A matrix gene ensembl ids vs miRNA family names. >=1 if interaction is predicted, 0 otherwise

Source

http://www.targetscan.org/vert_71/

targetscan predicted miRNA gene interactions

Description

targetscan predicted miRNA gene interactions

Usage

targetscan_symbol
targetscan_symbol

Usage

train_network_centralities
train_network_centralities

Format

(obtained by SPONGE method)

Package 'SPONGE'

Help Index

build classifiers for central genes

Description

Usage

Arguments

Value

tests and trains a model for a disease using a training and test data set (e.g., TCGA-BRCA and METABRIC)

Description

Usage

Arguments

Value

ceRNA interactions

Description

Usage

Format

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.

Description

Usage

Arguments

Value

Examples

Functions to define Sponge modules, created as all the first neighbors of the most central genes

Description

Usage

Arguments

Value

Calculate enrichment scores

Description

Usage

Arguments

Value

example potential central nodes

Description

Usage

Format

prepare ceRNA network and network centralities from SPONGE / SPONGEdb for spongEffects

Description

Usage

Arguments

Value

Function to calculate centrality scores Calculation of combined centrality scores as proposed by Del Rio et al. (2009)

Description

Usage

Arguments

Value

discretize #' (functions taken from: Jerby-Arnon et al. 2018)

Description

Usage

Arguments

Value

Computes an elastic net model

Description

Usage

Arguments

Value

Calibrate classification method

Description

Usage

Arguments

Value

Preprocessing ceRNA network

Description

Usage

Arguments

Value

Perform F test for gene-miRNA elastic net model

Description

Usage

Arguments

Value

Extract the model coefficients from an elastic net model

Description

Usage

Arguments

Value

Compute the residual sum of squares error for an elastic net model

Description

Usage

Arguments