Machine Learning Model and Molecular Docking for Screening Medicinal Plants as HIV-1 Reverse Transcriptase Inhibitors

Abstract

The human immunodeficiency virus type 1 reverse transcriptase (HIV-1 RT) plays a significant role in viral replication and is one of the targets for anti-HIV. However, a mutation in viral strains rapidly developed the resistance of the compounds to the protein, reducing the effectiveness of the inhibitors. This work seeks to utilize machine learning-based quantitative structureeactivity relationship (QSAR) analysis in combination with molecular docking simulations to forecast the presence of active compounds derived from medicinal plants. Specifically, the objective is to identify compounds that have the potential to operate as inhibitors of HIV-1 reverse transcriptase (RT), encompassing both wild-type and mutant variants. It is demonstrated that some substances are no longer suitable as inhibitors due to changes in the HIV-1 RT enzyme. Based on the screening results, four medicinal plants, Melissa officinalis, Punica granatum, Psidium guajava, and Curcuma longa, are worth further investigation. Nevertheless, the findings from the in vitro study suggest that extracts derived from pomegranate rind and guava leaves exhibit significant promise as HIV-1 RT inhibitors.