Artificial Intelligence Derived Artemisinin Drug Compound Acts as an Effective Candidate against SARS-CoV-2 Receptors: An In-Silico Study to Combat COVID-19

Abstract

Background: Global outburst of coronavirus has challenged the whole world to discover drugs to combat the current pandemic. Repurposing drugs is a promising approach as it provides new openings to challenge the emerging COVID-19. However, in the epoch of big data, artificial intelligence (AI) technology offers to leverage computational methods for finding new candidate drugs through an In-silico approach. Aim and Objectives: The aim and objectives of our present work basically are the designing of a plant-derived compound against the COVID-19 receptors which might act as effective therapy along with predicting the outcome of the disease with a deep learning program language that is python (anaconda) 2.7 version. Methodology: Artificial Intelligence technology helps in understanding the interactions of coronavirus with receptors through the computer-aided drug designing process (CADD). The ligand-protein interactions were prepared with the Maestro (Schrödinger) program which aids to study the docking pose of artemisinin compound with SARS-CoV-2 receptors like 7CTT, a nonstructural protein (NSP) and 7MY3 Spike glycoprotein. Thus, Artificial Intelligence technology examines the drug-target interaction with Neural Networking built with a deep learning machine algorithm and predicts the outcome of the disease with python program language. Results: Artemisinin exhibited the highest antiviral activity against the SARS-CoV-2 receptors like 7CTT and 7MY3. The three-dimensional structures of the ligands and SARS-CoV-2 receptors were retrieved from the PubChem Open Chemistry Database. The ligand-protein interactions were performed with the help of the Maestro (Schrödinger) program, which revealed MM/GBSA values of 7CTT interaction with derivative ligands of antimalarial compounds such as D95 (-45.424), artemisinin (-35.222), MPD (-31,021), MRD (-21.952) and 6FGC (-34.089), whereas with 7MY3 spike glycoprotein interactions MMGBSA values for D95 (-26.304), MPD(-18.658), MRD(-28.03) and 6FGC (-13.47) binding affinities have followed Lipinski rule of 5 and further predicted the outcome with random forest decision tree with an accuracy of about 75% with python program. Conclusion: Repurposing of the drug through an In-silico approach against the SARS-CoV-2 virus revealed its antiviral actions. The docking studies approach has shown the XP score, gliding energy, and MMGBSA values which were predicted with a deep learning program built with Artificial Intelligence technology.