Clusterpose
Short description of the program
From: Bob Diamond (MRC LMB, Cambridge)
Date: 25 January 2007
Clusterpose seeks families of similar structures within an ensemble of structures by repeatedly looking for the currently existing pair of clusters which may be superimposed on each other with a residual smaller than that for any other cluster-pair. Initially, each structure is a 'cluster' of 1. The resulting tree of combinations reveals groupings of similar structures. Three different criteria are available for measuring
the similarity of clusters, as described in Acta Cryst. (1995). D51, 127-135 and in the README file.
Alternatively, structures may be presented to Clusterpose in a sequence determined by the user, eg in energy order, and Clusterpose may be set to add these structures into a single cluster, one at a time, in the order given. This will show at what point the higher energy structures lose comparability with the low energy ones.
Clusterpose is a development of Polypose, which superimposes any number of structures in one operation. Polypose may also find multiple solutions if such exist. Both programs detect and report enantiomorphism, if present. See Protein Science (1992), 1, 1279-1287. Polypose is in CCP4.
Fortran command-line program which contains Clusterpose.
Python graphical interface around Fortran program.