Find in Library

In this study, novel potential inhibitors of SARS-CoV-2 variants were designed de novo using generative neural networks. The top-performing ligand based on docking performance and ADMET profiles was CID #526. It forms several hydrogen bonds with wild-type SARS-CoV-2, indicating its potential as an inhibitor of the receptor-binding domain. Mutated variants of the RBD also showed good interactions with CID #526, implying the inhibitory properties of our top-performing compound against various variants. Molecular dynamics analysis showed a stable ligand–RBD complex. CID #526 can easily be synthesized using low-cost starting molecules. Overall, the generated ligands merit further investigation to determine their efficacy and safety as a treatment for COVID-19.