ebdbNet (version 1.1)

visualize: Visualize Network Structure

Description

This function allows the user to visualize the network structure determined by the EBDBN, either by appropriately formatting results to be imported in Cytoscape, or invoking the Rgraphviz package to visualize network directly in R.

Usage

visualize(zscores, sig, format = "Cytoscape", type)

Arguments

zscores

The matrix of z scores obtained from ebdbn.

sig

The 0-1 matrix indicating absence or presence of an edge obtained from zCutoff.

format

Format of visualization

type

Type of network

Value

• edges
  If 'type' = "Cytoscape", a matrix of appropriately formatted values

Details

Allows visualization of the network structure, as determined by the z-scores from the posterior distribution of the network (calculated in ebdbn), and the 0-1 matrix identifying the absence or presence of an edge (calculated in zCutoff). The argument 'format' takes the value of "Rgraphviz" or "Cytoscape", and is used to determine how results should be formatted. For the former, the package Rgraphviz is invoked to draw a directed graph. For the latter, a table is created where the first column contains the identification number of "from" nodes, the second column contains the identification number of "to" nodes, and the third column contains the interaction type (1 = activation, -1 = inhibition). This table can be input into Cytoscape to visualize the network. The argument 'type' takes the value of feedback" or "input" depending on the type of network inferred.

Examples

library(ebdbNet)
tmp <- runif(1) ## Initialize random number generator
set.seed(125214) ## Save seed

## Simulate data
simData <- simFunc(R = 5, T = 10, P = 10, v = rep(10, 10), perc = 0.10)
Dtrue <- simData$Dtrue
y <- simData$y
 
####################################################
## Run EB-DBN without hidden states
####################################################
net <- ebdbn(input = "feedback", y, K = 0, 
	conv.1 = 0.15, conv.2 = 0.10, conv.3 = 0.10)

## Calculate sensitivities, specificities, and precisions of D matrix
z <- zCutoff(net$DPost, net$DvarPost)

## Create results matrix for Cytoscape, based on 95 perc significance
cytoscape.results <- visualize(z$z, z$z95, format = "Cytoscape", type = "feedback")

## Visualize using Rgraphviz, if appropriately installed
## library(Rgraphviz)
## visualize(z$z, z$z95, format = "Rgraphviz", type = "feedback")