Enabling Computational Exploration of an Expanding Academic Chemistry Space

Abstract Ultralarge virtual libraries have become a cornerstone of modern early stage drug discovery, enabling the rapid exploration of billions of compounds. The chemical space accessible to most screening campaigns, however, remains constrained not by the availability of building blocks, which are broadly shared across vendor catalogs, but by the narrow set of well-established reaction chemistries used to combine them. To address this limitation, we present the continued development of the Pan-Canadian Chemical Library (PCCL), a community-driven effort to expand the accessible chemical space beyond commercial libraries by incorporating synthetic reactions developed across academic laboratories in Canada. The current release of the PCCL is contributed by seven Canadian academic groups. Using broadly available commercial building blocks, the resulting reaction space spans on the order of 1017 theoretically reachable drug-like compounds. Moreover, using cheap building blocks from ChemSpace (≤\$100/100 mg), an explicitly enumerated and experimentally actionable sublibrary of approximately 77 M compounds suitable for high-throughput virtual screening was generated. Comparison with the Enamine REAL library reveals that PCCL occupies a largely orthogonal chemical space, with fewer than 0.1% Bemis–Murcko scaffolds shared between the two libraries. Rather than explicitly enumerating the full 1017 compound space, we introduce a PostgreSQL/RDKit-backed hit-expansion framework that enables targeted enumeration and analog retrieval directly within the reaction space. The expanded PCCL demonstrates how a community-sourced initiative can provide a scalable mechanism to access novel chemical space and support early stage drug discovery against challenging and underexplored biological targets.