This has occurred despite the advent of technologies such as QSAR, HTS and combinatorial chemistry. The difficulty in building effective, focused libraries and to filter or design libraries with desired properties has limited the success ratio of these approaches thus far.
To overcome these limitations, we have developed a series of datamining technologies based on the availability of large amounts of factual data. Such in-silico systems permit the seamless generation of focused libraries covering a wide range of chemical diversity patterns around specific mechanisms of action or selected chemical scaffolds. Furthermore, the approach has been expanded to measure ADMET properties.
As the technology can operate over a broad spectrum of chemical classes, handle molecular libraries of any size and adapt rapidly to new data, we have evaluated its applications in the natural product drug discovery field.
By selecting privileged scaffolds and fragments from natural sources and combining them following a series of transformation rules we have been able to construct focused molecular libraries in different therapeutic areas.
In addition to the generation of novel compounds, the system can characterize pharmacological and toxicological profiles of natural products for which the mechanism of action has not been elucidated.
As an example, the application of the technology has permitted to discover new multikinase inhibitors, derived from natural products magnolol and honokiol.