birtePredict: Prediction of gene expression via biRte.

In birte: Bayesian Inference of Regulatory Influence on Expression (biRte)

Description

Given a biRte model, this function makes posterior inference about gene expression data.

Usage

birtePredict(model, test.genes, method=c("Bayes", "MAP"), knock.out=NULL)

Arguments

model	output of birteRun
test.genes	Set of gene IDs. Gene IDs should be contained into the defined regulator-target gene network. Note that expression data is generally not required to be available for these genes.
method	Bayes: estimate expectation of posterior predictive distribution. MAP: Use previously fitted ridge regression model (birteFitRidge), which has to be stored into a slot "fit.ridge".
knock.out	optionally: A character vector of those regulators, which should be removed from the network before making predictions, hence simulating a knock.out

Value

#conditions x #replicates matrix containing data.frame objects with

gene	gene, for which predictions are made
mean	expected expression
sd	SD of predictions (only for method="Bayes")

Author(s)

Holger Froehlich

Examples

# artificial data
data(humanNetworkSimul)
sim = simulateData(affinities2)
limmamRNA = limmaAnalysis(sim$dat.mRNA, design=NULL, "treated - control")

# burnin and sampling size is much too small in reality
result = birteLimma(dat.mRNA=sim$dat.mRNA, data.regulators=NULL, 
limmamRNA=limmamRNA, 
affinities=affinities2, niter=100, nburnin=100, thin=2)

est = birtePredict(result, rownames(sim$dat.mRNA))

birte documentation built on May 2, 2019, 12:32 a.m.