generateLocalPPPs: Wrapper function to compute local pairwise posterior...
In uc-bd2k/gimmR: This package was built to cluster gene expression data using the Gaussian Infinite Mixture Model.

Description Usage Arguments Details Value Author(s) References See Also Examples

After running gimm and posthoc with the estimate_context option set to "y" and the intFiles option set to "TRUE", the user may specify a set of samples (called context) to compute the local pairwise posterior probabilities for all pairs of genes.

1
2

generateLocalPPPs(dataFile, M, T, samples = NULL, cc = NULL, out = NULL, wcS = NULL, nIter=1000, burnIn = 500, verbose = TRUE)
localGeneClustering(dataFile, out = NULL, wcS = NULL, cc = NULL, samples = NULL, burnIn = 500)

`dataFile`	file name used in previous gimm/posthoc run
`M`	total number of samples
`T`	total number of genes
`samples`	a set of samples. must be a subset of (1:M)
`cc`	if specified, a matrix of sample to context assignments where rows represent Gibbs steps and columns represent samples.
`out`	if specified, a matrix of global gene cluster assignments where rows represent Gibbs steps and columns represent genes.
`wcS`	if specified, a matrix of local gene cluster assignments where rows represent Gibbs steps as well as global clusters and columns represent samples.
`nIter`	nIter used in previous gimm/posthoc run.
`burnIn`	burnIn used in previous gimm/posthoc run.
`verbose`	want progress updates?

The function first creates a file similar to the .out file, then runs the C function posthoc. The output (i.e. the local PPPs) is stored in a .zm3 file.

An error code ('0' for success) produced when running the C function "posthoc."

Johannes Freudenberg

http://eh3.uc.edu/gimm

runGimmNPosthoc, computeDCEscore

data(GalData)
gimmOut <- runGimmNPosthoc(GalData, dataFile="galData", M=20, T=820, estimate_contexts="y", intFiles=TRUE, verbose=TRUE)
generateLocalPPPs(dataFile="galData", M=20, T=820, samples=which(cutree(gimmOut$hSClustData, k=2)==1), nIter=10000, burnIn=5000)
localPPPs <- read.table("galData.zm3")[, 1]
localPPPs <- matrix(localPPPs, 820, 820)
localGtr <- hclust(as.dist(1 - localPPPs), method = "average")
localGtr$labels <- 1:820
plot(localGtr)