Abstract

This poster describes a cheminformatics approach using KNIME [1] to design a screening library which represents diversity in fragment chemistry [2] and geometry as well as considers chemical features amenable to X-ray crystallography. This novel approach searches across commercially available fragment-like chemical space to identify diverse compounds in terms of chemical features, conformational space and molecular features pertaining to 3D-shape. Library diversity is enhanced through the measure of our ‘Cocktail Score’ as the algorithm optimally searches commercial chemical space to finally produce a library which is enriched with 3D-like structures determined by thresholds in the sum of the normalized principal moments of inertia (NPR1+NPR2) and the plane of best-fit (PBF).