i have performed a time-course microarray experiment, found differentially expressed genes, and now intend to cluster those by their expression pattern over the time.
maSigPro provides a way for hierarchical clustering via "hclust" method.
however the number of clusters that it forms is not deriven from the data but authomatically set to 9. It also is possible to define the number with "k". I thought the idea of hierarchical clustering is that the result is estimated from the data?
Is there something that i miss?
An alternative approach would be to use "kmeans" clustering with a k determined by a v-Fold Cross-Validation.
Are there any easy to use packages that could perform it, that you would suggest?