Why Semisynthetic Accessibility Should Influence Which Natural Product Scaffolds Advance from Computational Prioritization to Medicinal Chemistry Development

Download PDF

Computational prioritization can identify natural products with attractive predicted binding, bioactivity, selectivity, or phenotypic signatures, but it does not determine whether medicinal chemistry can interrogate those scaffolds efficiently after nomination. This article develops an original decision framework in which semisynthetic accessibility qualifies, rather than replaces, pharmacological priority. The analysis distinguishes molecular synthetic complexity, retrosynthetic feasibility, locally available reactions and starting materials, parent or precursor supply, chemoselective modifiability, stereochemical obligations, and the experimentally reachable analogue neighborhood. These dimensions answer different questions and should not be collapsed into a single notion of synthesizability. The central proposed contribution is chemical learnability: the practical ability to generate structurally discriminating analogues that test structure–activity, selectivity, physicochemical, and liability hypotheses while preserving the features that made a natural-product scaffold interesting. Semisynthesis is positioned as one route to such learning, alongside late-stage functionalization, biosynthetic engineering, chemoenzymatic synthesis, and total synthesis. The framework assigns no universal weights, thresholds, or pass–fail rules, and chemically difficult scaffolds are not presumed inferior. Exceptional pharmacology, engineered supply, or new selective chemistry may justify continued advancement. Conversely, a highly ranked scaffold may warrant deprioritization when accessible chemistry cannot interrogate the relevant structure. Semisynthetic accessibility is therefore proposed as a context-dependent advancement variable whose value lies in estimating what medicinal chemistry can realistically learn next, not in predicting therapeutic success.