FMOPhore for hotspot identification and efficient fragment-to-lead growth strategies

Fragment based drug design (FBDD) is like a chess game in that a good or a bad move can dramatically influence the outcome. At the start of the design process, it is important to identify the key binding site residues (hotspots) that can have a substantial impact on ligand efficiency (LE) and binding. Here, we introduce a novel, fully automated algorithm named FMOPhore, which performs Quantum Mechanics Fragment Molecular Orbital (QM-FMO) calculations on 3D-protein-ligand pharmacophore models. This is implemented in a novel scoring function named FP-score to classify binding site residues in two classes: 1) Hotspot residues (further delineated into three categories; Anchor, Transient, and Accessible) and 2) Non-hotspot residues. We apply our algorithm in two different scenarios: holo-complex and apo-structure scenarios, testing its robustness on sixteen different protein targets including the application for fragment growing and target selectivity. We demonstrate that handling protein binding site flexibility using Dy-FMOPhore improves hotspots detection. The FMOPhore algorithm can be a powerful tool in identifying and quantifying binding site hotspots to enable an efficient design strategy for fragment-to-lead optimization.