Definition of Known Drug Space - Analysing Known Drug Space using the Potential Energy Surface as a Possible Molecular Descriptor

Abstract

Known drug space is among the most investigated types of the common chemical spaces and is increasingly generating considerable interest in terms of the drug discovery and development fields. Mainly a neglected area in this field is the dire need to develop and establish adequate approaches that will facilitate the computation of known drug properties and determine the required areas of chemical space which can then be employed in the design of drugs. Accordingly, the aim of this research was to evaluate the applicability of using a number of minima in the potential energy surface maps of known drugs as a theoretical approach and as a molecular descriptor in defining known drug space. The data for 1451 known drugs were gathered from DrugBank, Chemspider and PubChem databases. Calculations of the minima in the potential energy surface maps were accomplished using Scigress software package. Then, the data analysis was performed using Microsoft Excel and the correlations between the number of minima, the rotatable bonds and the ring counts were identified. According to these tests, the most marked observation to emerge from the data comparison was the significant positive (direct) correlation between the minima counts and the number of rotatable bonds in the drug molecules with (R² = 0.968) which provided a new theoretical basis for using rotatable bonds as a main molecular descriptor for known drug space. Moreover, this correlation has contributed to predict a range of minima where 68.4% of the sample (992 drug compounds) occupied the range from 0 to 20 minima in their potential energy surface maps. In the second set of the data analysis there was no remarkable correlation observed between the minima counts and the number of rings in the drug compounds, nevertheless it has provided further evidence that the developability of drug molecules is strongly connected to the lower number of aromatic rings contained in the compound (0 to 3 aromatic rings). Although, molecular descriptors in general should not be treated as rigid filters and should be given the essential priority beside other criteria, we assume that this work could be the framework for establishing a new molecular descriptor and facilitating further definitions of known drug space.