Figure 2.

Pipeline for docking-based, “two-way” virtual screening strategy. GOLD software was used to dock the drug library against each of the 48 predicted binding sites. All docking scores were converted to fit quality (FQ) scores. Subsequently, the FQ scores of each drug were transformed into Z scores for each site. Out of the 1,268 FDA-approved drugs, 190 were prioritized with a Z score ≥ 2 at 5 or fewer target sites; these were considered high confidence hits. Consequently, 17 overlapped with a set of drugs previously known to have marginal to high potency against SARS-CoV-2 in vitro and were considered virtual hits¹⁰. These were then subjected to MD simulations, in which 6 drugs were observed to bind stably to their respective target protein. A medium confidence set of drugs was also prepared based on a Z score ≥ 1.8, which prioritized 276 drugs; 33 of these overlapped with the reference drugs. Overall, there were 34 and 97 unique drug-site pairs for the high- and medium confidence hits respectively that overlapped with the reference drugs. In-silico, prioritized drugs are shown in cylinders with solid lines and reference drugs with in vitro activity against SARS-CoV-2 in cylinders with dashed-dotted lines. Prepared using PowerPoint Presentation Software.