Fig. 4: Our own N-layered Integrated molecular Orbital and molecular Mechanics (ONIOM) scheme with machine learning potentials (MLPs), and electron density maps. | Nature Communications

Fig. 4: Our own N-layered Integrated molecular Orbital and molecular Mechanics (ONIOM) scheme with machine learning potentials (MLPs), and electron density maps.

From: Accelerating reliable multiscale quantum refinement of protein–drug systems enabled by machine learning

Fig. 4

a Chemical structures of imatinib, CPI-0610 (with the ONIOM(MLP-CC:MLP-DFT) scheme) and osimertinib. Structures for the binding sites of (b) imatinib in spleen tyrosine kinase (PDB ID: 1XBB), (c) CPI-0610 in bromodomain (PDB ID: 5HLS) and (d) osimertinib in EGFR (PDB ID: 6JX4) from various quantum refinements using M7, M8 (M9a), M10 (M10a) and the X-ray crystal structures, including the electron density maps (2mFo-DFc maps, contoured at 1.0 σ (blue), mFo-DFc maps, contoured at +3.0 σ (green), and mFo-DFc maps, contoured at –3.0 σ (red)). The X-ray results were taken from the experimental structures without further refinement. MLP-CC (coupled-cluster quality of MLP); MLP-DFT (density functional theory quality of MLP).

Back to article page