From data to knowledge-Use of the Cambridge Structural Database for studying molecular interactions

The ordered arrangement of molecules and ions in crystal structures is the result of non-bonded interactions, albeit modulated in the crystalline state by random crystal packing forces. Each individual crystal structure gives but a snapshot of these interactions. It is only by examining a large number of structures containing similar molecules or functional groups that we can gain any generalized knowledge about the rules governing molecular interactions. Indeed if we knew these rules it should be possible to predict crystal structures likely to be formed by specific molecules, whereas at present each new crystal structure requires a de now experimental solution. We would also be able to evaluate the relative contribution of non-bonded and crystal packing forces and extrapolate to the interactive behaviour of molecules in solution and, most importantly, in the environments encountered in biological systems. A key to these studies is the Cambridge Structural Database (CSD) system where experimental results from over 100, 000 individual structure determinations are stored in computer readable form. The system also provides the computational tools needed to locate the relevant structures and analyse the numerical data using a variety of statistical techniques. This paper describes the latest Version of the CSD system (Version 5) released Autumn 1992. The use of the system is illustrated by several examples, such as the preferential interaction of certain functional groups, the study of the C-H center dot center dot center dot O bond and drug-DNA interactions.