Using a distance matrix in evalclusters Linkage for Gap estimation

Hi,

When using evalclusters with 'Linkage' clustering methods and 'Gap' estimation, is it possible to give a distance matrix instead of a data matrix as input? I have my own distance computing function and so have already computed the distance matrix. Using Linkage by itself lets you use a distance matrix as input but evalclusters i don't know. It works, but then the optimal number of clusters is very different than when i use evalclusters with the original data. (Which by the way is a collection of signals and i want to cluster them).