Novel Development of Predictive Feature Fingerprints to Identify Chemistry-Based Features for the Effective Drug Design of SARS-CoV-2 Target Antagonists and Inhibitors Using Machine Learning

Abstract

A unique approach to bioactivity and chemical data curation coupled with random forest analyses has led to a series of target-specific and cross-validated predictive feature fingerprints (PFF) that have high predictability across multiple therapeutic targets and disease stages involved in the severe acute respiratory syndrome due to coronavirus 2 (SARS-CoV-2)-induced COVID-19 pandemic, which include plasma kallikrein, human immunodeficiency virus (HIV)-protease, nonstructural protein (NSP)­5, NSP12, Janus kinase (JAK) family, and AT-1. The approach was highly accurate in determining the matched target for the different compound sets and suggests that the models could be used for virtual screening of target-specific compound libraries. The curation-modeling process was successfully applied to a SARS-CoV-2 phenotypic screen and could be used for predictive bioactivity estimation and prioritization for clinical trial selection; virtual screening of drug libraries for the repurposing of drug molecules; and analysis and direction of proprietary data sets.

Artificial Intelligence Technologies for COVID-19 De Novo Drug Design

“…Molecules that showed more than 75% inhibition were considered active, and among the wide range of FDA-approved drugs, 267 were identified as active. Leveraging the bioactivity data of ChEMBL and a subset of the data, it was trained on physicochemical characteristics from the set of 110 chemical properties, and a set of 868 compounds was then used to establish a binary classification model capable of predicting whether a molecule was active or inactive in the test; overall the model was able to identify the bioactivity with an accuracy of 65% [ 47 ].…”

Artificial Intelligence Technologies for COVID-19 De Novo Drug Design

“…Molecules that showed more than 75% inhibition were considered active, and among the wide range of FDA-approved drugs, 267 were identified as active. Leveraging the bioactivity data of ChEMBL and a subset of the data, it was trained on physicochemical characteristics from the set of 110 chemical properties, and a set of 868 compounds was then used to establish a binary classification model capable of predicting whether a molecule was active or inactive in the test; overall the model was able to identify the bioactivity with an accuracy of 65% [46]. E. Glaab et al reported a combined virtual screening study, molecular dynamics (MD) simulation, machine learning, and in vitro experimental validation analysis, which led to the identification of small molecule inhibitors of 3CLpro with micromolar activity and to a pharmacophore model.…”

Machine learning in handling disease outbreaks: a comprehensive review

“…This review study identified 71 articles that covered the field. Bioinformatics applications have been developed for discovering, repositioning and repurposing drugs to find effective clinical treatments [55], [56]. The studies analyzed the interactions between compounds of drugs and proteins of SARS-CoV-2 and predicted the bio-activities that occurred.…”