Automatization of the bottom-up exploration of the Chemical Space for the early drug discovery

Designing new medicines more efficiently and cost-effectively is closely linked to the exploration of chemical space. On-demand chemical libraries are expanding every year, aiming to include as much of the vastness of chemical space as possible. Currently, the size of these collections surpasses the throughput capabilities of standard high-throughput virtual screening (HTVS), requiring intelligent solutions rather than brute-force docking. In response, we have developed a new platform for mining these extensive libraries, capable of rapidly screening billions of compounds to identify those with the highest probability of binding to specific therapeutic targets and possessing favorable ADMET properties. To demonstrate proof of concept, we have tested this approach on relevant oncogenic targets and experimentally validated the selected compounds through biological assays with therapeutically relevant readouts. The automatization of this complicated protocol into a platform is necessary to provide universal, access to state-of-the-art drug discovery capabilities to researchers anywhere. This will incentivize the exploration of novel targets and mechanisms of action that are perceived as too risky or insufficiently lucrative.