Graph Convolutional Network-Based Screening Strategy for Rapid Identification of SARS-CoV-2 Cell-Entry Inhibitors

Area of Science:

• Virology
• Computational Chemistry
• Drug Discovery

Background:

• SARS-CoV-2 cell entry is a key target for antiviral drug development.
• Heparan sulfate proteoglycan (HSPG) and actin are crucial for viral entry.
• Conventional drug screening is time-consuming and limited in scope.

Purpose of the Study:

• To develop an efficient virtual screening method for identifying SARS-CoV-2 entry inhibitors.
• To leverage graph convolutional networks (GCNs) for drug repurposing.
• To accelerate the discovery of novel small molecule inhibitors.

Main Methods:

• Trained a GCN-based classification model using known HSPG and actin inhibitors.
• Virtually screened 170,000 compounds to identify potential drug candidates.
• Validated hits through cell-based assays using SARS-CoV-2 pseudotyped particles.

Main Results:

• Identified approximately 2000 potential hits from virtual screening.
• Shortlisted 256 active compounds using HSPG cargo uptake assays.
• Discovered 16 compounds with significant inhibitory activity against SARS-CoV-2 entry.

Conclusions:

• A GCN-based virtual screening workflow enables rapid identification of inhibitors for validated drug targets.
• This AI-driven approach significantly accelerates the discovery of novel antiviral small molecules.
• The identified compounds represent promising candidates for further development against SARS-CoV-2.