Design and Analysis of Diverse Screening Libraries

Screening very similar compounds is wasteful. The paradigm that structurally similar molecules exhibit similar physicochemical and biological properties (1) can be used to reduce existing screening pools to small representative subsets which enhances the efficiency to find novel lead structures. The stategy is applied to two examples: a combinatorial library based on a single scaffold and an existing structurally diverse compound database. While the combinatorial library contains 97 % redundant structures, for the compound database 62 % of the structures are redundant and can be filtered out leading to optimal diverse libraries. The obtained diverse subsets enhance the structural diversity of the Maybridge database.