Mpropred: A machine learning (ML) driven Web-App for bioactivity prediction of SARS-CoV-2 main protease (Mpro) antagonists

• DOI: 10.1371/journal.pone.0287179
• Publication Date: 2023-06-24T20:56:35Z
• Authors: Nadim Ferdous; Mahjerin Nasrin Reza; Mohammad Uzzal Ho...; Shahin Mahmud; Suhami Napis; Kamal Chowdhury; A. K. M. Mohiuddin
• Keywords: Matthews correlation coefficient; Polyproteins; Molecular descriptor

The severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) pandemic emerged in 2019 and still requiring treatments with fast clinical translatability. Frequent occurrence of mutations spike glycoprotein SARS-CoV-2 led the consideration an alternative therapeutic target to combat ongoing pandemic. main protease (M pro ) is such attractive drug due its importance maturating several polyproteins during replication process. In present study, we used a classification structure–activity relationship (CSAR) model find substructures that leads anti-M activities among 758 non-redundant compounds. A set 12 fingerprints were describe M inhibitors, random forest approach was build prediction models from 100 distinct data splits. set’s modelability (MODI index) found be robust, value 0.79 above 0.65 threshold. accuracy (89%), sensitivity specificity (73%), Matthews correlation coefficient (79%) calculate performance, also statistically robust. An extensive analysis top significant descriptors unveiled significance methyl side chains, aromatic ring halogen groups for inhibition. Finally, predictive made publicly accessible as web-app named pred order allow users predict bioactivity compounds against . Later, CMNPD, marine compound database screened by our app all results revealed their binding affinity Molecular dynamics (MD) simulation molecular mechanics/Poisson Boltzmann surface area (MM/PBSA) showed improved properties complexes. Thus, knowledge shown herein can develop more effective specific inhibitors accessed https://share.streamlit.io/nadimfrds/mpropred/Mpropred_app.py