GA_search_connected: Genetic algorithm for module identification on weighted...

Description Usage Arguments Value Author(s) References Examples

Description

Using genetic algorithm (GA) to identify active modules on weighted network. The connectedness of resulted module is ensured by connected components finding of binary encoded set of genes. Rewritten from COSINE package.

Usage

1
2
GA_search_connected(lambda, node_score, edge_score, EdgeList, num_iter = 1000,
  muCh = 0.05, zToR = 10, minsize = 30)

Arguments

lambda

weight parameter in the objective

node_score

A N-length vector storing the nodes

edge_score

A M-length vector storing the edges

EdgeList

A two-columns matrix with each row a pair of weighted edges

num_iter

Number of iterations in GA

muCh

Mutation rate in GA

zToR

Zero to one ratio

minsize

The minimal size of module

Value

a list containing module size, best score, module as a list of nodes and a list of GA functions.

Author(s)

Dong Li, dxl466@cs.bham.ac.uk

References

Ma, Haisu, et al. "COSINE: COndition-SpecIfic sub-NEtwork identification using a global optimization method." Bioinformatics 27.9 (2011): 1290-1298.

Dong Li et al. Memetic algorithm for finding active connected subnetworks in intracellular networks. 2016.

Examples

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
library(COSINE)
data(scaled_node_score)
data(scaled_edge_score)
data(PPI)
GA_result <- GA_search_connected(lambda=0.5,scaled_node_score,
scaled_edge_score,PPI,num_iter=100, muCh=0.05, zToR=10, minsize=50)
GA_result2<-GA_search_PPI(lambda=0.5,scaled_node_score,scaled_edge_score,PPI, 
num_iter=100, muCh=0.05, zToR=10, minsize=50)
## visualized by igraph
selected = c()
for (i in 1:dim(PPI)[1]){
    if (PPI[i,1] %in% GA_result$Subnet || PPI[i,2] %in% GA_result$Subnet)
        selected = c(selected,i)
}
library(igraph)
g <- graph.data.frame(PPI[selected,], directed=FALSE)
V(g)$color <- "gray"
V(g)$color[match(GA_result$Subnet,V(g)$name)] <- "red"
layout <- layout.reingold.tilford(g, circular=T)
png(filename = 'Connected_PPI.png', width=4096, height = 3652)
plot(g,layout=layout, vertex.size=5, vertex.label.cex=5,vertex.label.dist=0.5,edge.width=5)
dev.off()

fairmiracle/PPINet documentation built on May 16, 2019, 9:59 a.m.