Design of Virtual Combinatorial Libraries

Abstract

Last year marked the completion of the sequencing effort on the human genome. Advances in the fields of genomics and bioinformatics are widely expected to bring forth a large number of new biochemical targets for drug design by linking specific diseases to single genes or collections thereof. The pharmaceutical and biotechnology sector is now likely to become one of the most active industrial fields in the new century. A dramatic increase in the number of pharmaceutical targets in the near future would create a bottleneck of sorts at the stage of pharmaceutical drug discovery. Historically, this field has used a serial process, screening and optimizing compounds one at a time. However, the advent of combinatorial chemistry and high-throughput screening has significantly altered the face of drug discovery. Rapid generation and screening of combinatorial libraries has become commonplace in both industry and academia; with the ever-growing offering of organic reagents along with the vast palette of organic reactions, the chemical space accessible to combinatorial chemists has dramatically expanded. The application of combinatorial technology, however, is not without a potential caveat—even as the cost of synthesis and testing of a single chemical entity has fallen, it skyrockets when multiplied by the thousands or millions of combinatorial library members. Some of the potential solutions include (i) use of the three-dimensional target structure information in the design process; (ii) application of pharmacophore and quantitative structure-activity relationship (QSAR) methods to focus on libraries that most closely resemble screening hits; and (iii) reduction in the scale on which both medicinal chemistry and screening are currently performed. These solutions imply a need for a rational approach to reducing library size, custom izing the library for a given target, and more efficient in silico screening of a library (or, potentially, multiple libraries) of compounds against a large number of targets of interest. The challenge in accomplishing these goals is in being able to reduce the library while increasing the potential hit rate of individual library members.