Topic automatically created for discussing the designs at:
https://covid.postera.ai/covid/submissions/b6889685-5fe5-4b2c-9cfd-330c9ed47202
@mc-robinson I know John Chodera is working on docking. Will there be some sort of docking screening for the submitted compounds? I’m asking as I can generate more compounds (with more interesting shapes including more of the most likely ‘warheads’) in a similar way to these compounds. The challenge I have is filtering the potential hits further.
Less important - I tried to select all of the fragments but could not submit. Error message something about being longer than 50 characters. I’m sure this is low on your priority list but just a mention as this design used all of the fragments as inspiration.
Thank you for putting this all together and even more so for keeping it going!
Hi @Waztom, sorry about that issue with the fragments, I guess I limited the character length in the database. Will take that into account.
I know @JohnChodera, @RGlen, and several others are working on docking. @rajatkpal has also previously reached out if people needed help on that front I believe. If you post your expansion ideas in the asking for help section, perhaps we Get Help section, perhaps someone could help try them out.
I’ve been docking the noncovalent inhibitors, but we’ll need to identify another way to select the covalent inhibitor design ideas that have been sketched since it’s much harder to model the covalent docking!
Hi, we’re still (Jason at CCDC) porting GOLD to our HPC system so we can basically parallel doc. We should be able to dock and score early next week I hope, There are a few issues we also are addressing wrt the crystal structures, Gerard Bricogne at Global Phasing is kindly re-refining the published structure from the ED, this hopefully will inform us of for making some changes to the orientation/pKa and tautomers of the histidines and some of the other AAs. It’s very difficult to ‘see’ hydrogen in x-ray and these are inferred from the structure. We need to be sure we have a decent model of this (at physiological pH) before doing all the calculations. An example is H163, which is in the binding site, and is critical to a few of the interactions seen in ligands for this class of proteases. Automated hydrogen addition can be problematic.
3 Likes
Refined X-ray structures with thoughtfully-assigned protonation and tautomeric states would be of immense value! Can you share those structures publicly when you’ve produced them?
For reference, our prepped receptors (for OpenEye FRED docking) and docked structures are currently being collected here:
Documentation is poor right now, but will be fleshed out over the weekend as we get the free energy calculations for the Moonshot compounds spun up.
1 Like
@JohnChodera Thank you so much! I just just had a look at the repo. Looks like we’re getting somewhere compared to the first round of submissions! What docking score (hybrid2) would get you really excited or is it more a combo of docking metrics? Learning as I go here.
@Daren_Fearon any chance that can we help with John’s request re ‘Refined X-ray structures with thoughtfully-assigned protonation and tautomeric states’ above?
We are about to get very busy in the lab again this week collecting data for follow-ups/new COVID-19 projects and it sounds like @rglen and Gerard are already on this so we’ll leave it in their hands (a benefit of sharing all of the data).