Mapping of protein surface cavities and prediction of enzyme class by a self-organizing neural network

An automated computer-based method for mapping of protein surface cavities was developed and applied to a set of 176 metalloproteinases containing zin c cations in their active sites. With very few exceptions, the cavity searc h routine detected the active site among the five largest cavities and prod uced reasonable active site surfaces, Cavities were described by means of s olvent-accessible surface patches. For a given protein, these patches were calculated in three steps: (i) definition of cavity atoms forming surface c avities by a grid-based technique; (ii) generation of solvent accessible su rfaces; (iii) assignment of an accessibility value and a generalized atom t ype to each surface point. Topological correlation vectors were generated f rom the set of surface points forming the cavities, and projected onto the plane by a self-organizing network. The resulting map of 865 enzyme cavitie s displays clusters of active sites that are clearly separated from the oth er cavities, It is demonstrated that both fully automated recognition of ac tive sites, and prediction of enzyme class can be performed for novel prote in structures at high accuracy.