Abstract

Primary high-throughput screening of commercially available small molecules collections often results in hit compounds with unfavorable ADME / Tox properties and low IP potential. These issues are addressed empirically at follow-up lead development and optimization stages. In this work, we describe a rational approach to the optimization of hit compounds discovered during screening of a kinase focused library against abl tyrosine kinase. The optimization strategy involved application of modern chemoinformatics techniques, such as automatic bioisosteric transformation of the initial hits, efficient solution-phase combinatorial synthesis, and advanced methods of knowledge-based libraries design.

Keywords: high-throughput screening, combinatorial synthesis, virtual screening, tyrosine kinase inhibitors, hit optimization, bioisosteric approach, quantitative structure-activity relationship, support vector machines