Hi
I have created pharmacophores based on the 44 crystal structures of covalent binders and then used these pharmacophores to screen all the covalent ideas submitted to the covid_submissions_03_31_2020 excel file.
I have used Schrodingers Phase tool to carry out the pharmacophore hypothesis generation and searching.
There were 742 covalent ideas submitted, the pharmacophore search was run on 731 of these compounds, 13 compounds were removed as they failed a structure check during preparation for the pharmacophore search.
The 3D sdf containing the 3D structures, CID, PhaseScreenScore and Phase_HypoID properties can be found in the sdf attachment.
An EXCEL file containing smiles strings, CID, PhaseScreenScore and Phase_HypoID properties is also attached.
PhaseScreenScore is the pharmacophore scoring result for the idea.
The highest scoring pharmacophore match is retained.
The higher the score, the better the fit to the pharmacophore hypothesis.
I would recommend scores at least >2 as being a minimum cutoff for closest mapping of the idea to the original fragment crystal structure.
There are about 35 ideas with a score of >2.5.
Phase_HypoID indicates which fragment crystal structure the idea maps best to.
For example if an idea has x1311_Hypothesis in the Phase_HypoID column, then it is matching the pharmacophore generated from the x1311 crystal structure.
covid_covalent_310329_phase_screen_1-hits_SCORED_3D.sdf (1.9 MB)
covid_submissions_03_31_2020_Pharmacophore_Results.xlsx (47.8 KB)