One of my student has recently finished a Studienprojekt, where he implemented
a tool to manipulate and study clusters of matches. This is essentially a GUI
where you can choose different strategies for clustering matches: clustering 
based on positions of the matches (cf. the paper by Salzberg and Volfovsky) or 
on sequence content (i.e.\ the matches have global alignment with few errros).
Clusters can be selected, and the matches in a cluster are 
isplayed as a match graph like in genalyzer.
