interventionalDBN-package: Interventional Inference for Dynamic Bayesian Networks

In interventionalDBN: Interventional Inference for Dynamic Bayesian Networks

Description

This package allows a dynamic Bayesian network to be inferred from microarray timecourse data with interventions (inhibitors).

Details

Package:	interventionalDBN
Type:	Package
Version:	1.2.2
Date:	2014-01-03
License:	GPL version 2 or newer
LazyLoad:	yes

Includes functions for formating the data (formatData), estimating the effects of an intervention (interventionEffects) and performing network inference (interventionalInference).

Author(s)

Simon Spencer

Maintainer: Simon Spencer <[e-mail]>

See Also

interventionalInferenceAdvanced, countGraphs, interventionalData, linesROC, nxt, trueMatrix, warshall.

Examples

library(interventionalDBN)
data(interventionalData)# loads interventionalData.
# Load your own data spreadsheet using myData<-read.csv("myDataFile.csv").

# Estimate nodes downstream of intervention.
egfriEffects<-interventionEffects(interventionalData,1,"DMSO","EGFRi")
aktiEffects <-interventionEffects(interventionalData,1,"DMSO","AKTi")

# Format the data for network inference
d<-formatData(interventionalData)

# EGFRi is active in conditions 2 and 4, AKTi is active in conditions 3 and 4.
# Each condition has 8 timepoints.
Z<-matrix(0,32,15)
Z[9:16,1]<-1 # EGFR (node 1) is inhibited in condition 2
Z[25:32,1]<-1 # EGFR (node 1) is inhibited in condition 4
Z[17:24,8]<-1 # AKT (node 8) is inhibited in condition 3
Z[25:32,8]<-1 # AKT (node 8) is inhibited in condition 4

# Perform network inference
myNetwork<-interventionalInference(d$y,d$X0,d$X1,Z,max.indeg=3,
  perfectOut=TRUE,fixedEffectOut=TRUE)

# Make ROC curve, to see how well we have done.
data(trueMatrix)
plot(0:1,0:1,t="l",col="grey",xlab="False positive rate",ylab="False negative rate",
  main="ROC curve showing network inference performance.")
redArea<-linesROC(trueMatrix,myNetwork$pep) # ROC area is also sent to the console.

# More realistically, the true edge matrix is unknown.
# We can use descendancy to get (a much coarser) ROC,
# which is based only on nodes that are downstream of the inhibitors.
pap<-warshall(myNetwork$pep)
effectMatrix<-matrix(NA,15,15)
effectMatrix[1,]<-1*(egfriEffects$p.values<=0.1)
effectMatrix[8,]<-1*( aktiEffects$p.values<=0.1)
blueArea<-linesROC(effectMatrix,myNetwork$pep,col="blue")
legend("bottomright",c("Edge matrix known","Descendancy ROC"),col=c("red","blue"),lty=1)

Example output

  Stimulus EGF : 8 / 8  baseline observations used.
  Stimulus EGF : 8 / 8 inhibited observations used.
  Stimulus EGF : 8 / 8  baseline observations used.
  Stimulus EGF : 8 / 8 inhibited observations used.
n = 32 
4 conditions:
  Cell line 1 : 32 samples.
  Inhibitor DMSO : 8 samples.
  Inhibitor EGFRi : 8 samples.
  Inhibitor AKTi : 8 samples.
  Inhibitor EGFRi+AKTi : 8 samples.
  Stimulus EGF : 32 samples.
  Time 0 : 4 samples.
  Time 1 : 4 samples.
  Time 2 : 4 samples.
  Time 3 : 4 samples.
  Time 4 : 4 samples.
  Time 5 : 4 samples.
  Time 6 : 4 samples.
  Time 7 : 4 samples.
n = 32 , nodes = 15 
a = 2 
Processing 576 models Tue Jan 29 07:52:44 2019 
Estimated duration 0.002873474 minutes.
Actual duration 0.002172311 minutes.
Renormalising...
Calculating posterior edge probabilities...
red ROC area = 0.7266314 
blue ROC area = 0.8888889 

interventionalDBN documentation built on May 2, 2019, 4:04 p.m.