Electronic-Structure Informatics for Drug Development

Drug development involves intricate chemical processes with versatile information across the dimensional hierarchy. Herein, electronic-structure informatics (ESI), informatics at the electronic-structure level, is discussed with particular emphasis on the unique definition of descriptors for machine learning. Since the time-independent Schrödinger equation describes the information at this level, three ESI methodologies can be suggested: (i) Hamiltonian-based ESI (H-ESI), (ii) wavefunction-based ESI (Ψ-ESI), and (iii) energy-based ESI (E-ESI). Among them, the E-ESI approach, which applies the unique and theoretically justified descriptors in machine learning, and its role in drug development are explained in detail in this chapter. Some applications of the E-ESI approach are also introduced in order to demonstrate the usefulness of the E-ESI method.