Importance of Structural Databases, Molecular Pharmacophores, Supramolecular Heterosynthons, and Artificial Intelligence–Machine Learning–Neural Network Tools in Drug Discovery

The progress and growth of drug discovery and development (DDD) in the past five decades are reviewed in terms of the changing trends over the years. The importance of the Cambridge Structural Database (CSD) and the Protein Data Bank (PDB) starting in the 1990s brought in the phase of structure-based drug design (SBDD). The supramolecular synthon led to the heterosynthon, which became the cornerstone for crystal engineering of multicomponent cocrystals and salts (MCCS) as improved medicines. Numerous studies on enhancing the solubility and permeability of biopharmaceutics classification system (BCS) class II and IV drugs in the decades of 2000–2020 resulted in a paradigm shift toward supramolecular crystalline complexes as drug substances, namely, MCCS instead of molecule-based drugs, new chemical entity (NCE), or new molecular entity (NME) entries. With the numerical explosion in the number of possible druglike substances and their pharmaceutical cocrystals and salts as improved materials, artificial intelligence (AI), machine learning (ML), and neural networks (NN) were introduced as computational tools to accelerate drug discovery decision making. This review ends with a thought on integrating the abovementioned advances over the past three decades to propose a hierarchic model for DDD with varying levels of difficulty and complexity for success in different resource settings. With over a million crystal structures in the CSD and over 200 000 protein structures in the PDB, together with cheminformatics tools for prediction, synthesis, and crystallization, integrated drug discovery is poised for rapid advances in the future.