Chemical libraries underpin contemporary drug discovery by enabling systematic exploration of chemical space for ligands. This review summarizes practical principles for designing, comparing, and selecting libraries, emphasizing how “diversity” can be quantified and engineered. Approaches to describe diversity include scaffold‐based analyses, measures of 3D character (for example, the fraction of sp 3 ‐hybridized carbon atoms), fingerprint‐based similarity comparisons, and shape‐ or descriptor‐based chemical space visualizations. Strategies to introduce diversity are discussed across building blocks, functional groups, stereochemistry, and scaffolds, including diversity‐oriented synthesis, biology‐oriented synthesis, and late‐stage skeletal editing to access related cores from common intermediates. We compare untagged plate‐based collections used in high‐throughput screening including diversity collections, drug repurposing sets, target‐focused libraries, and fragment libraries, with affinity‐selection formats that identify binders without individual activity assays. For affinity selection, we cover DNA‐encoded libraries decoded by polymerase chain reaction and next‐generation sequencing, and mass‐spectrometry‐based formats such as peptide nucleic acid‐encoded, peptide‐encoded, and self‐encoded libraries. Collectively, these concepts provide a framework to align library type, synthesis constraints, and screening readout with the biological question.
Hansen et al. (Tue,) studied this question.