I have a data set with similarity scores (0-1) determined by two different methods - that is, I have two similarity scores for each pair of items in a set. These are analogous to sequence similarities (or the corresponding distances) in the sense that there is no real coordinate space involved, only distances between pairs. I would like try biclustering these, but all I've found so far, searching through various documents about biclustering, are methods that appear to be based on the actual coordinates.

Can anyone point me to any kind of distance-only bi- or mult-clustering? Thanks in advance for any consideration.