ABSTRACT
We present some new ideas for characterizing and comparing large chemical databases. The comparison of the contents of large databases is not trivial since it implies pairwise comparison of hundreds of thousands of compounds. We have developed methods for categorizing compounds into groups or series based on their ring-system content, using precalculated structure-based hashcodes. Two large databases can then be compared by simply comparing their hashcode tables. Furthermore, the number of distinct ring-system combinations can be used as an indicator of database diversity. We also present an independent technique for diversity assessment called the saturation diversity approach. This method is based on picking as many mutually dissimilar compounds as possible from a database or a subset thereof. We show that both methods yield similar results. Since the two methods measure very different properties, this probably says more about the properties of the databases studied than about the methods.
