Regression modeling and virtual screening of natural products exhibiting antibacterial activity. an application of electronic-structure informatics descriptors

Abstract

Machine learning using descriptors suggested for "electronic- structure informatics (ESI)"is carried out to establish a regression model for antibacterial activity of 60 natural products. The obtained model is used for the in-house data containing 175 molecules. The screening result is justified through literature survey in which, at least, 13 compounds listed in the top 20 compounds are shown to have the antibacterial activity.