is there a proper way to choose clusters in hierarchical clustering?

Question

[v.berriosfarias]

Hello, I'm just trying to determine the number of clusters to look into a dendogram so I can make a proper co assembly of paired end reads (metagenomic dataset):

is there any R package to estimate the proper number of clusters that indicates the branches to cluster ?

I'm using a distance matrix, so hclust()package seemed to be a good clustering method due to the fact that I could not find a partitioning clustering method that accept a distance matrix as input.

Thanks for reading :)

[seidel]

Do you have a definition of "proper"? What properties would a good set of clusters have for you? How would you know if had the right number of clusters?

[Hamid Ghaedi]

Using consensus clustering approach , you would be able to find the optimal number of cluster based on some statistics like CDf plot . Even using this kind of method you have to make sure that the number you are going to pick as optimal number of cluster, have to has biological meaning as well. You did not clariifed to what context your work is related to, want to get more helpful comments , provide much more details on the problem that you have.