Accelerated Hit Detection in Combinatorial Chemistry Libraries.

CHIRON can be used for combinatorial chemistry projects to efficiently perform hit/lead optimization. This implies that CHIRON finds the most active molecule of a combinatorial library without the necessity to actually synthesize and analyze the large majority of all molecules (more than 80%). The molecule search is driven by a genetic algorithm, which selects more active molecules from generation to generation. For the first generation, CHIRON chooses some molecules from the specified combinatorial library on bases of structural diversity. All following generations, being made up of the same number of molecules as the first generation, are selected on bases of the biological activity of the molecules from previous generations. This work-flow iterates until an optimum is reached.

The biological activity, which drives the algorithm, can either come from experimental data as for example from a kinetic screen or can be determined by computational techniques. Intelligent Pharma offers both ligand-based and receptor-based virtual biological activity measurements, which are performed using HELIOS or SELENE, respectively. This allows for a completely automated combinatorial chemistry porject using huge molecule libraries, which molecule content can, if desired, be completely selected by the client. In this case CHIRON determines the most active compounds, being the best hits from within the molecule library, which chemical structures are provided to the customers.

At Intelligent Pharma the performance of CHIRON was tested using three different experimentally analyzed combinatorial libraries bearing around 15000 molecules each. Experimentally determined biological activities were used as the driving force of the molecule search. CHIRON finds the mayority of biologically most active molecules from a combinatorial library after choosing a subset of 8 to 13% of all its members for synthesis and analysis (statistics sent upon request at info@intelligentpharma.com). As combinatorial libraries are synthesized in an intelligent manner directed towards minimizing the number of synthesis steps, CHIRON would save at least 65% and at most 90% of synthesis work in a spatially addressed synthesis. Furthermore, 87 to 92% of biological screens are avoided, resulting in a tremendous reduction in both work and costs.