Quantum-level machine learning is playing a crucial role in predicting the potential energy surface (PES) of molecular torsions in drug molecules, a key factor in drug design. A recent study used the ANI1x neural network potential to predict the PES of the antiparkinsonian drug molecule, Selegiline. The study successfully calculated the vibrational frequencies, electronic energy, and optimization of the molecular structure of Selegiline in a short computing time, suggesting high efficiency for computational structure-based drug design studies. Machine learning is revolutionizing quantum chemical calculations, accelerating drug discovery processes, and is expected to lead to the development of more effective and safer drugs.
↧