Combined use of computational chemistry and chemoinformatics methods for chemical discovery
Data analysis on numerical data by the computational chemistry calculations is carried out to obtain knowledge information of molecules. A molecular database is developed to systematically store chemical, electronic-structure, and knowledge-based information. The database is used to find molecules related to a keyword of “cancer”. Then the electronic-structure calculations are performed to quantitatively evaluate quantum chemical similarity of the molecules. Among the 377 compounds registered in the database, 24 molecules are found to be “cancer”-related. This set of molecules includes both carcinogens and anticancer drugs. The quantum chemical similarity analysis, which is carried out by using numerical results of the density-functional theory calculations, shows that, when some energy spectra are referred to, carcinogens are reasonably distinguished from the anticancer drugs. Therefore these spectral properties are considered of as important measures for classification.