CLUSTER-ANALYSIS OF THE BIOLOGICAL MACROMOLECULE CRYSTALLIZATION DATABASE

Cluster analysis was performed on the Biological Macromolecule Crystallization Database (BMCD) [1] in an effort to uncover trends useful in the crystallization of new macromolecules. The following crystallization parameters were used in defining an experiment; pH, temperature, molecular weight, macromolecular concentration, precipitant type and crystallization method. Using these parameters, a measure of the difference between experiments is developed. Groups or clusters of similar experiments are identified as those close together based upon the difference measure. Descriptive statistics were performed on each cluster. Since each cluster represents a well defined class of macromolecules, it seems reasonable to use the boundaries of conditions within each cluster as initial conditions when attempting to crystallize a new macromolecule.