evalnetworksNames: Computing all the statistics from data of biological network

In netmes: Exploratory Analysis of Gene Networks with Netmes, An R package for network-based measures

Description

This function takes the adjacency matrix, the data sets folder path and number of mutual information steps for the therhold mutual information. It returns averaged mutual information matrix, averaged mutual information matrix for the inferred edges, the averaged mutual information values for the inferred edges, the averaged mutual information values for the non-inferred edges, maximum F-scores, threshold mutual information values that gives maximum F-score, and also the same input adjacency matrix of genes.

Usage

evalnetworksNames(E, net, infilepath, namesSif)

Arguments

E	number of experiments.
net	Adjacency matrix of genes (not symmetric).
infilepath	The path of the directory, where all the data files kept.
namesSif	names of the genes corresponding to the adjacency matrix.

Details

It reads all the mutual information matrices from the folders and take the average and then derive the other parameters from it. This function is designed for biological networks where the gene names need to be derived along with the gene interactions from the files.

Value

evalnetworksNames returns an object of class '"res"' which is a list with components:

aveMImat	averaged mutual information matrix of the overall experiments.
rmat	averaged mutual information matrix for the inferred edges of the overall experiments.
tpre	All the averaged mutual information values for the inferred edges.
fnre	All the averaged mutual information values for the non-inferred edges.
Fscores	Maximum F-score values for all the experiments.
Ithr	The threshold mutual information values, which gives maximum F-score, for all the experiments.
net	Adjacency matrix of genes (not symmetric).

References

Gokmen Altay and Frank Emmert-Streib. Revealing differences in gene network inference algorithms on the network level by ensemble methods. Bioinformatics, 2010, 26(14),1738-1744.

Frank Emmert-Streib, Gokmen Altay. Local network-based measures to assess the inferability of different regulatory networks. IET Syst. Biol. 2010, Volume 4, Issue 4, p.277-288.

See Also

readSifNet, readNet, evalnetworksNames

Examples

# E <- 5 # number of experiments

# infilenet <- "./netmes/data/networkEcoli.sif"
# infilepath <- "./netmes/data/bio/"
# This is an example path. Change this path wrt the path of data in your computer.

# resEC <- readSifNet(infilenet, infilepath)   # true network (+1 edge/ activator - -1 edges/repr.
# net <- resEC$net

# res <- evalnetworksNames(E, resEC$net, infilepath, resEC$names)

# boxplot(res$Fscores)

# save(res, file = "res_EColi")
# if desired for later analysis.

netmes documentation built on May 2, 2019, 5:21 p.m.