Autonomous Phenotype-Directed Molecular Discovery

The challenges associated with developing new medicines are very significant indeed. The process generally involves the selection of a protein target that is associated with a particular disease; the identification and optimisation of molecules that can modulate that target; and the execution of clinical trials involving a nominated drug candidate. The cost of bringing each new drug to the market is over £2 bn, in large part because of crippling (>95%) attrition rates in the drug discovery/development process. Even when successful, the process typically takes about 12 years from laboratory to patient. The pharmaceutical sector therefore faces the major challenges of increasing both productivity (by reducing costs and time to patient) and innovation (by finding drugs with new modes of action and/or for new disease areas).

Bioactive small molecules can play crucial roles throughout the drug discovery process. Before a specific protein is selected as a drug discovery target, it is crucial to establish its role in both health and disease. In these fundamental biomedical science studies, chemical probes can serve as enabling tools that can allow the cellular function of proteins to be established. Historically, the biology of proteins has been explored very unevenly, and biologists' favourite proteins of study are largely unchanged in >20 years; however, it has been shown that the availability of chemical probes can have a transformative effect on which proteins are the subject of biomedical investigations. Once a protein target has been chosen, it is necessary to discover a safe and effective drug that allows a specific medical need to be addressed. Over 90% of prescribed medicines, and well over half of newly-approved drugs, are small molecules.

The discovery of bioactive small molecules is currently very resource-intensive, contributing to around 20% of overall drug discovery costs (including the costs of the majority of programmes that ultimately fail). Molecular discovery is generally driven by cycles in which candidate molecules are designed, prepared and evaluated. This approach generally enables series of molecules to be investigated one-by-one. To expedite discovery, a narrow toolkit of reliable chemical reactions is used, which tends to limit the diversity of the molecules that are evaluated.

Through this Centre-to-Centre collaboration, we will develop a new approach for discovering bioactive small molecules. The discovery process will be driven by the functional ("phenotypic") effect of the molecules that are prepared. It will contrast starkly with overwhelming current practice in which specific series of molecules are designed ahead of preparation and evaluation. We will deliberately harness reactions that have been developed at the University of Leeds (UoL) that will enable hundreds of diverse molecules to be prepared in parallel. The phenotypic effect of these molecules will be determined at the Max Planck Institute for Molecular Physiology (MPIMP). The reactions that yield functional products will inform the design of hundreds of further reactions using algorithms developed at RFI. The approach will enable the autonomous search for functional small molecules, and it has the potential to transform both medicinal chemistry (drug discovery) and chemical biology (biomedical science).

The ambitious goal to realise autonomous phenotype-directed molecular discovery will only be possible by integrating the unique features of the collaborating Centres: fragment/diversity-oriented synthesis (UoL), high-throughput experimentation (RFI), algoithms (RFI), phenotypic screening (MPIMP), chemical proteomics (UoL) and specific biomedical science (MPIMP). To demonstrate the value of the new approach to end-users, we will demonstrate that it can enable the discovery of novel bioactive molecules, for example chemical probes that can provide new insights into biological mechanisms that underpin disease and/or healthy ageing.