The sample data that I have is labeled into few classes. When I run the coefmap command on the ExpressionMix / blood sample, the samples are shuffled by classification. Thats great of course. But I would like to have the same plot but with the samples unshuffled (without the hirarchical clustering step). Or ist there an alternative? Here's my current code:

eMix = ExpressionMix(eset)
res <- gedBlood(eMix, verbose = TRUE)
coefmap(res,annCol=labels)
basismap(res,subsetRow=FALSE)