This design focused on exploring the chemical space constrained by the typical functional scaffolds found in the fragment hits.
Viewing in Fragalysis - overlaying all the fragments yields regions where some functional groups agglomerate. The shape of the fragment is likely impacted by the functional groups. Some functional group scaffolds have a larger impact on the shape than others. Procedure: 1. BRICS algorithm to find synthetic building blocks of the fragment hits 2. BRICS build potential compounds from building blocks 3. Filter potential compounds via Lipinski rule 4. Predict binding via Random forest classification model trained on fragment hits/misses *Docking will very likely do this bit better and/or filter down to potentially better hits *Need to update fragment data with several new fragment addtions