gaga (version 2.18.0)

parest: Parameter estimates and posterior probabilities of differential expression for GaGa and MiGaGa model

Description

Obtains parameter estimates and posterior probabilities of differential expression after a GaGa or MiGaGa model has been fit with the function fitGG.

Usage

parest(gg.fit, x, groups, burnin, alpha=.05)

Arguments

gg.fit

GaGa or MiGaGa fit (object of type gagafit, as returned by fitGG).

x

ExpressionSet, exprSet, data frame or matrix containing the gene expression measurements used to fit the model.

groups

If x is of type ExpressionSet or exprSet, groups should be the name of the column in pData(x) with the groups that one wishes to compare. If x is a matrix or a data frame, groups should be a vector indicating to which group each column in x corresponds to.

burnin

Number of MCMC samples to discard. Ignored if gg.fit was fit with the option method=='EBayes'.

alpha

If gg.fit was fit with the option method=='Bayes', parest also computes 1-alpha posterior credibility intervals.

Value

An object of class gagafit, with components:

parest

Hyper-parameter estimates.

mcmc

Object of class mcmc with posterior draws for hyper-parameters. Only returned if method=='Bayes'.

lhood

For method=='Bayes' it is the posterior mean of the log-likelihood. For method=='EBayes' it is the log-likelihood evaluated at the maximum.

nclust

Number of clusters.

patterns

Object of class gagahyp indicating which hypotheses (expression patterns) were tested.

pp

Matrix with posterior probabilities of differential expression for each gene. Genes are in rows and expression patterns are in columns (e.g. for 2 hypotheses, 1st column is the probability of the null hypothesis and 2nd column for the alternative).

Details

If gg.fit was fit via MCMC posterior sampling (option method=='Bayes'), parest discards the first burnin iterations and uses the rest to obtain point estimates and credibility intervals for the hyper-parameters. To compute posterior probabilities of differential expression the hyper-parameters are fixed to their estimated value, i.e. not averaged over MCMC iterations.

References

Rossell D. GaGa: a simple and flexible hierarchical model for microarray data analysis. http://rosselldavid.googlepages.com.

See Also

fitGG to fit a GaGa or MiGaGa model, findgenes to find differentially expressed genes and posmeansGG to obtain posterior expected expression values. classpred performs class prediction.

Examples

#Not run
#library(EBarrays); data(gould)
#x <- log(exprs(gould)[,-1])  #exclude 1st array
#groups <- pData(gould)[-1,1]
#patterns <- rbind(rep(0,3),c(0,0,1),c(0,1,1),0:2) #4 hypothesis
#gg <- fitGG(x,groups,patterns,method='EBayes')
#gg
#gg <- parest(gg,x,groups)
#gg