A friend of mine was interested in selecting stable genes from RNA-Seq experiment, stable meaning genes with expression not affected by changing conditions.
I think the easiest solution is using partitioning function implemented in cummeRbund (K-means clustering):

ic<-csCluster(myGenes,k=4)
head(ic$cluster)
icp<-csClusterPlot(ic)
icp

Of course, it’s good to play with `k` value, so you indeed get some cluster with genes that are of interest.
Then, once you have some candidates, you may want to look for genes having similar expression patterns.

mySimilar<-findSimilar(cuff,"my_gene_name",n=20)
mySimilar.expression<-expressionPlot(mySimilar,logMode=T,showErrorbars=F)
mySimilar.expression