Cortex's AI discovers potent antivirals for COVID-19

Our team trained our AI using an array of publicly available data quantifying the effectiveness of compounds at inhibiting SARS-CoV-2. Guided by the neural network’s predictions, we selected six compounds from a pool of drugs previously approved for human trials in unrelated conditions, none of which had been explored for their potential against COVID-19.

These molecules - that we selected entirely virtually - were then tested against the real live virus in the lab of Dr. Olagnier, in Aarhus University, Denmark. Remarkably, five out of the six selected compounds demonstrated antiviral effects at non-toxic doses, as predicted by our neural network.

Among these, two candidates have further shown impressive efficacy against various SARS-CoV-2 variants across diverse cell types, including human lung cells in 3D cultures. These compounds have exhibited both preventive and curative capabilities, proving effective when administered before and after cell exposure to the virus.

Two more compounds where also confirmed active by the lab of Pr. Shirit Enav - head of a virology lab at Stanford University, California - using a Luciferase-based cellular assay.

For more information, check out our publication in Antiviral Research!