FrAncestor A Multi-Dimensional Toolkit for Fragment Library Design and Optimisation

The success of fragment-based drug discovery is closely tied to the quality and diversity of the fragment library used. This paper introduces FrAncestor, a novel toolkit developed to design high quality fragment screening libraries by efficiently mining commercially available chemical space. FrAncestor integrates multiple molecular dimensions including physicochemical properties, molecular sociability, pharmacophore diversity, and reaction vectors (RVs) within a Molecular Genealogy Tree (MGT) framework. It supports the design of reversible, covalent, target-focused or general-purpose libraries. The workflow analyses subsets of chemical space from GDB-11, GDB-13, GDB-17, FDB-17, ChEMBL-35, Zinc-fragment databases, Enamine Real Space (HAC 11–23), commercially available vendor libraries, and fragment structures derived from the Protein Data Bank (PDB-fragments). FrAncestor was used to construct optimised fragment libraries for the Drug Discovery Unit (DDU), offering a practical strategy for informed library design.