Quantum machine learning methodology for predicting the inhibitory activity of drugs against COVID-19

Abstract

The study investigated the potential use of a quantum machine learning algorithm to forecast molecules ability to inhibit the SARS-CoV-2 coronavirus. An experimental dataset comprising 1904 previously identified compounds was utilized in order to accomplish this. The structure of each compound was then used to determine four molecular descriptions, which were then used as input data for the quantum learning algorithm. The experimental inhibitory activity was then used as a label to categorize the compounds as either "Active" or "Inactive." Using this knowledge, a four qubit quantum training algorithm was created. These preliminary results demonstrate the potential of using quantum computations to find possible molecules that could be candidates in the fight against the coronavirus, as the obtained result had a 95 % accuracy rate.