attract (version 1.24.0)

calcInform: Function calculates the informativeness metric (average MSS) for a set of cluster assignments.

Description

Function calculates the informativeness metric (average MSS) for a set of cluster assignments.

Usage

calcInform(exprs.dat, cl, class.vector)

Arguments

exprs.dat

a matrix of gene expression values.

cl

a vector of cluster assignments.

class.vector

a vector specifying the group membership of the samples.

Value

A numeric value representing the average MSS value (informativeness metric) for a set of cluster assignments. For an informative cluster, the RSS values should be very small relative to those produced by the informativeness metric (the MSS values).

Details

This function is also called internally by findSynexprs.

References

Mar, J., C. Wells, and J. Quackenbush, Defining an Informativeness Metric for Clustering Gene Expression Data. to appear, 2010.

Examples

## Not run: 
# library(cluster)
# data(subset.loring.eset)
# clustObj <- agnes(as.dist(1-t(cor(exprs(subset.loring.eset)))))
# cinform.vals <- NULL 
# for( i in 1:10 ){
# 	cinform.vals <- c(cinform.vals, calcInform(exprs(subset.loring.eset), cutree(clustObj,i), pData(subset.loring.eset)$celltype))
# }
# k <- (1:10)[cinform.vals==max(cinform.vals)]			# gives the optimal number of clusters 
# ## End(Not run)