Topic automatically created for discussing the designs at:
https://covid.postera.ai/covid/submissions/MAR-TRE-143fb005
We counter-screened these PPIs against a panel of human cystein proteases with formation of a covalent bond as a constraint.
This graphic has the molecules (not in submission order) in rows and the proteins in columns with the first column being 5RF7. It is colour coded: White not a hit; yellow moderate score; amber good score; red very good score.
It is apparent that there is scope to select molecules which are predicted to be highly selective unlike noncovalent cysteine protease inhibitors. Some of the proteins appear to be much more promiscuous than others but this promiscuity is different from the non-covalent bonding discussed separately. Drawing any conclusions from merely one set of covalent and one set of non-covalent potential inhibitors is probably not statistically sound but this approach does provide a rational approach to selecting molecules to purchase or make and test which is preferrable to a highly subjective or random approach!