Researchers at Neurocrine Biosciences were interested in discovering novel SNRI ligands. Working with a small training set of known inhibitors, they were able to generate many structures similar to patented SNRI ligands using an early version of the de novo design engine now implemented in Muse. They reasoned that because known SNRI structures unrelated to the training set were rediscovered in the de novo design process, there was a good chance that some of the novel structures generated might also be active hits. This in fact was the case resulting in the identification of novel Intellectual Property (IP) for Neurocrine. The authors stated: “Most importantly, one of the top scoring compounds (generated) was found to be highly active and was selected as lead compound in the project at Neurocrine.”
Researchers at Neurocrine Biosciences had identified a promising and novel scaffold for GnRH receptor ligands. They were interested in exploring novel R-group substituents in two positions in order to generate a broad range of intellectual property. Using an early version of de novo design now in implemented and enhanced in Muse, Neurocrine scientists generated over 655 side chain ideas that met their design criteria. The majority of these appeared to be reasonable to the chemists. After minor modifications to facilitate synthesis, thirteen molecules were synthesized and tested. These structures incorporated completely novel side chains not represented in the training set of GnRH ligands. Several potent and novel structures were found, expanding the IP space of the novel scaffold to include novel side chains that would have been missed.