find_localControllingFeatures: Find local controlling features

In CoNI: Correlation Guided Network Integration (CoNI)

Description

This function applies for a selected subnetwork a binomial test using the frequency of appearance of an edge feature and the total number of edge features. The probability corresponds to 1/n_df, where n_df corresponds to the total number of edge features in the network. The selected subnetwork corresponds to the second level neighborhood of a specific node. The test is applied to all possible second level neighborhoods in the network.

Usage

find_localControllingFeatures(ResultsCoNI, network, padjust = TRUE)

Arguments

ResultsCoNI	The output of CoNI (after p-adjustment)
network	Network created with the function generate_network
padjust	logical. Filter output based on adjusted p values

Value

Returns a data.frame with the results of the binomial tests. Significant results correspond to local controlling features

Examples

#Load color nodes table
data(MetColorTable)

#Assign colors according to "Class" column
MetColorTable<-assign_colorsAnnotation(MetColorTable)

#Load CoNI results
data(CoNIResultsHFDToy)

#Generate Network
#Note: Colors not visible when ploting in Igraph
HFDNetwork<-generate_network(ResultsCoNI = CoNIResultsHFDToy,
                             colorVertexNetwork = TRUE,
                             colorVertexTable = MetColorTable,
                             Class = MetColorTable,
                             outputDir = "./",
                             outputFileName = "HFD",
                             saveFiles = FALSE)

#Note: For this tiny example nothing is significant
LCG_BinomialTestTableHFD<- find_localControllingFeatures(ResultsCoNI = CoNIResultsHFDToy,
                                                         network = HFDNetwork )
LCGenes_HFD<-as.character(unique(LCG_BinomialTestTableHFD$edgeFeatures))

CoNI documentation built on Sept. 30, 2021, 5:09 p.m.