Solutions for Computational Chemist
Molecular interaction field (MIF), aiming at describing molecules as their ability to interact with any chemical entity, and related molecular descriptors are one of the most established and versatile concepts in drug discovery. Clearly, any further improvement in the in silico molecular description is highly desirable for novel and more reliable drug design and medicinal chemistry applications. Here we applied a hybrid quantum mechanics and machine learning approach to better parameterize the hydrogen-bond potentials of small molecules which allowed us to derive more chemical-aware force fields and MIFs. Finally, we derived new molecular and atomic descriptors, that will be available in VolSurf3, and we will show examples of their usability in drug design and medicinal chemistry applications.