Machine intelligence design of 2019-nCoV drugs

Kaifu Gao; Duc Duy Nguyen; Rui Wang; Guo-Wei Wei

The partition coefficient, which measures how hydrophilic or hydrophobic a chemical substance is, is defined as the ratio of concentrations of a solute in a mixture of two immersible solvents at equilibrium. 33 The logarithm of partition coefficient, denoted logP, is a well-known coefficient which plays an essential role in governing kinetic and dynamic aspects of drug action. In this paper, we will employ an Open-Source Cheminformatics Software Rdkit 22 to calculate logP values of our 2019-nCoV drug candidates to evaluate the reliability of the potential 2019-nCoV drugs we predicted. All of the logP values of all predicted molecules can be found in the Supplementary Materials. While the logP values of the predicted top 15 drug candidates are presented in Table 2 . From the table, it can be observed that most 2019-nCoV drug candidates we predicted have the logP value smaller than 5, which matches one of the rules in "Lipinski's rule of five". 34 Moreover, the ritonavir, an HIV protease inhibitor already on the market, has a predicted logP = 5.91, which shows that our potential drugs with logP values slightly larger than 5 can still be considered as druggable molecules.