README.md

In QiliShi/NetworkSim: Network Comparison Based on Structural Equivalence and Graphlets

NetworkSim

Current version:1.1.1, Depends: R (>= 3.4.0)

Description

NetworkSim, an R package that provides two different ways for undirected network comparison, which contains the structural equivalence based on paired nodes and the graphlet-based systematic measurement.

Installation

NetworkSim currently needs to be installed using devtools. The igraph and orca packages also need to be installed.

devtools::install_github("QiliShi/NetworkSim")

Running times

Create different sizes of PPI networks

library(NetworkSim)
library(igraph)

## 
## Attaching package: 'igraph'

## The following objects are masked from 'package:stats':
## 
##     decompose, spectrum

## The following object is masked from 'package:base':
## 
##     union

small_net1 <-subnet(databases.net,nodes = c(names(neighbors(databases.net,'TP53')),'TP53')[1:75],neighbors = F)
small_net2 <-subnet(databases.net,nodes = c(names(neighbors(databases.net,'TP53')),'TP53')[40:120],neighbors = F)

middle_net1 <-subnet(databases.net,nodes = c(names(neighbors(databases.net,'TP53')),'TP53')[1:300],neighbors = F)
middle_net2 <-subnet(databases.net,nodes = c(names(neighbors(databases.net,'TP53')),'TP53')[150:450],neighbors = F)

nodes <- c(names(neighbors(databases.net,'TP53')),'TP53',names(neighbors(databases.net,'MYC')),'MYC',
           names(neighbors(databases.net,'APP')))
large_net1 <-subnet(databases.net,nodes = nodes[1:2000],neighbors = F)
large_net2 <-subnet(databases.net,nodes = nodes[500:3300],neighbors = F)

nodes <- names(V(databases.net))
Vlarge_net1 <-subnet(databases.net,nodes = sample(nodes,15000),neighbors = F)
Vlarge_net2 <-subnet(databases.net,nodes = sample(nodes,15000),neighbors = F)

Measure the running times on small (tens of nodes, hundreds of edges), medium (hundreds of nodes, thousands of edges), large ( thousands of nodes, tens of thousands of edges), and very large ( tens of thousands of nodes, hundreds of thousands of edges) sizes of PPI networks from Human Protein Reference Database (HPRD), BioGRID and mentha databases.

system.time(paired <- nodesCom(middle_net1,middle_net2))

##    user  system elapsed 
##    0.90    0.01    1.12

system.time(netSE(middle_net1,middle_net2,paired[1,1],paired[1,2]))

##    user  system elapsed 
##    0.01    0.00    0.01

system.time(netODA(middle_net1,middle_net2))

##    user  system elapsed 
##    1.82    0.02    2.34

Examples

Create the input networks with igraph types

1.Create networks from our intergrated human PPI network.

TCGA_AF_2687.net<-subnet(net=databases.net, nodes=TCGA_AF_2687.Muts)

2.Create networks from user's own data, which should be a data frame containing a symbolic edge list in the two columns. For example:

library(sand)

## Loading required package: igraphdata

## 
## Statistical Analysis of Network Data with R
## Type in C2 (+ENTER) to start with Chapter 2.

g <- graph.data.frame(elist.lazega,directed=F)

Get the paired nodes between two networks

paired<-nodesCom(TCGA_AF_2687.net, TCGA_A6_2686.net)

Calculate a similarity of a pair of nodes

netSE(TCGA_AF_2687.net, TCGA_A6_2686.net, 'NMT1', 'MED10')

## [1] 0.1825742

Calculate all structural equivalence of the paired nodes

paired.SE<-netSEs(TCGA_AF_2687.net,TCGA_A6_2686.net, paired)

The overall agreement based on graphlets of two networks can be gained by

netODA(TCGA_AF_2687.net, TCGA_A6_2686.net)

## [1] 0.4195262

QiliShi/NetworkSim documentation built on July 23, 2019, 12:46 a.m.