Open-Source Coronavirus Monthly Update Meeting #3: Monday 11th July 2022.

[TomkUCL]

Today's monthly SARS-CoV-2 Helicase FBDD progress meeting will be taking place on Zoom. See link below to join. The video recording of today's meeting will be uploaded here shortly thereafter.

Join Zoom Meeting https://ucl.zoom.us/j/97172937586

Meeting ID: 971 7293 7586

To tag everyone involved in thi project on Github use @StructuralGenomicsConsortium/cnp4-nsp13

[mattodd]

Hi @TomkUCL - could you please add in some agenda items to the above? I think we need a quick summary of modelling/predictions from @kipUNC, a protein update from the Toronto crew (Tom, I don't think Peter Brown is on Github, but could be reached by mail), and then a chemical update that includes your "tracking sheet". A summary of what's needed next in the helicase review, and how people can contribute to the writing. @tmw20653 would you be able to say something about READDI-AViDD, too?

I've a new, last-minute family clash this evening. I'll try to join while on the move, but if I'm not there, start/proceed without me, and assign me any tasks that ought to be mine. Within reason.

All - remember this is a high-level overview of what's been going on, that should be understandable by @jamesday100. What's happened in the lasy month, and what do we need to do in the next month.

Remember also that, as things stand, this project is concerned with the potentially allosteric site of Nsp13, not the ATP site, though there will of course be some overlaps there in terms of proteins/assays/compound design methods. The ATP site is the focus on the current CACHE challenge (https://cache-challenge.org/competitions/competition-2). Tom's going to post some images that clarify the relationship between the different sites.

[TomkUCL]

[TomkUCL]

@kipUNC @StructuralGenomicsConsortium/chemistry

Here are the quotes from Enamine in Excel:

Copy of Quote_1546888_EUR_Enamine.xlsx

Copy of Quote_1546887_EUR_Generated.xlsx

You will see that both quotes include ‘Custom Synthesis’ items. The Custom Synthesis items are relatively expensive starting from ~300-500 euros per sample.

Could you please inspect the quotes and let me know if there are any Custom Synthesis compounds which you would like to receive an estimation for?

I can also visually inspect these in DataWarrior (files attached below) to get an idea of how tricky these look to make. For the Custom Synthesis molecules, we need a way to prioritise these since there are a lot. We could either do this by Glide score (start with the lowest scoring compounds) or as @jamesday100 suggested, dock and visually inspect them based on likely key interactions; I am liaising with Geoff Wells here at UCL SoP who has suggested PyRx to screen on my own machine, or alternatively has some personal protocols for docking & rescoring with Vina and Autodock-GPU that can use multiple protein structures from MD on Linux machines they also streamline some of the output formatting.

[TomkUCL]

Thanks to everyone from @StructuralGenomicsConsortium/cnp4-nsp13 and project pharma champion @jamesday100 who attended yesterday's monthly nsp13 meeting. The meeting video link can be found here:

https://ucl.zoom.us/rec/play/hV4ToBK_wwRRBiBvopNsmeRCO1RV7i6sWN5sU0_pFWKsxRaPFKAVLX4NPB8NSPz0uZiutjkz4W9tXUgt.6mwdX9k9OQ7_ty8z Passcode: .Ax?8+.O

This list has been submitted to Enamine and we are currently waiting to hear back with a quote. @drc007 @edwintse Do you recommend any other supplier databases that are worth checking?

[TomkUCL]

Slide deck from yesterday's meeting: nsp13 monthly update no.3.pptx

[TomkUCL]

The nsp13 collected hits list (250 small molecules) sent from @kipUNC and discussed in the meeting can be found here: nsp13_collected_hits.xlsx

I have begun cluster analysis of these structures in DataWarrior, starting with the 'Generated' list (100 molecules).. Top 100, generated.zip

...based on neighbour structure similarity:

...based on Glide score:

...and with the seven purchasable mRDB compounds from the Enamine price quote coloured dark green, the spread looks like this:

And the same for the 'Enamine' list (150 molecules).. Top 250, enamine.zip

...based on neighbour structure similarity:

...based on Glide score:

...and with the fifty-three purchasable mRDB compounds from the Enamine price quote coloured dark green, the spread looks like this

...and for both Enamine and generated lists combined... Generated and Enamine 350 combined - Copy.zip

...based on Glide score:

....based on neighbour structure similarity, where there are some obvious clusters:

[drc007]

Thanks to everyone from @StructuralGenomicsConsortium/cnp4-nsp13 and project pharma champion @jamesday100 who attended yesterday's monthly nsp13 meeting. The meeting video link can be found here:

https://ucl.zoom.us/rec/play/hV4ToBK_wwRRBiBvopNsmeRCO1RV7i6sWN5sU0_pFWKsxRaPFKAVLX4NPB8NSPz0uZiutjkz4W9tXUgt.6mwdX9k9OQ7_ty8z Passcode: .Ax?8+.O

This list has been submitted to Enamine and we are currently waiting to hear back with a quote. @drc007 @edwintse Do you recommend any other supplier databases that are worth checking?

Enamine is a good start, Mcule https://ultimate.mcule.com and Wuxi Galaxi https://www.biosolveit.de/wp-content/uploads/2021/08/Xu.pdf could be worth looking at also.

[TomkUCL]

@StructuralGenomicsConsortium/cnp4 here is the makeable compound list from enamine (~60 molecules):

Copy of Quote_1548556_EUR.xlsx

[TomkUCL]

@StructuralGenomicsConsortium/cnp4-nsp13

We now have 60 purchasable compounds from Enamine, leaving ~190 compounds to be made. We can request custom synthesis for the remaining 190 for around EUR 300-500 per molecule if any look particularly promising.

Based on the above DataWarrior analysis, what are people's thoughts on how to proceed with the synthesis of non-purchasable compounds?

Option 1) @mattodd and I think that prioritising based on Glide score is the way to go.

Option 2) we prioritise structural diversity by picking exemplar molecules from different clusters - based on the similarity of the Glide scores, this would help to broaden the chemical space.

Option 3) we dock and manually inspect and prioritise likely favourable interactions (slower).

@jamesday100 do you have a bias here?

[tmw20653]

@TomkUCL @mattodd @ahsgc We could pursue both Option 1 and 2. Option 1: you pick the top scoring Glide hits to synthesize at UCL Option 2: we send exemplars from each cluster for synthesis at our CRO. Can you generate a list for @ahsgc to review?

[jamesday100]

@TomkUCL @mattodd @ahsgc

Sounds like the preferred option is 1 and 2.

In my experience docking scores have very little correlation with binding affinity and it's much better to use SBDD knowledge, shape fit to rank and prioritize hit matter. We are not dealing with large numbers here only ~100 compounds or so. I guess I do not know your access to tools to dock and inspect the binding modes. And ranking by glide maybe an easier and better method.

Normally with any virtual screening it’s done on commercial libraries, a modeler would inspect the binding modes and conformations to check they have a chance of working and create a prioritized list for purchase.

With generated compounds the energy barrier to obtaining them is higher as you have to work out a synthesis and that can sometimes take alot of time. So the time upfront working out their probability of success is often quicker than the time it takes to synthetize the compounds. Hence you have a time saving and a greater chance of success.

Other ways you could filter these compounds:

Size/MW

Undesirable functional groups (I saw an N-oxide in one of the picture above)

An option to consider guys is just to buy the 60 compounds from Enamine, screen and see what you get? This will give us an idea of hit rate etc.

Cheers,

James

[kipUNC]

@jamesday100

Those compounds are not the result of a "blind" docking/VS, those are produced with knowledge of SBDD, p-phore modeling etc. Each of that top one hundred are not just top but score, they were also examined to fit of SBDD data and interaction "maps". Also I would like to point out that for this site SBDD only produced very few (3/4 don't recall of he top of my head) quote similar ligands.

Saying all that I think it is a great idea if somebody would be interested to look at the 3D poses in the site. I just remember that as the number of available (at reasonable cost) compounds is not huge Tim suggested just to buy it.

[mattodd]

Helicases examined as part of the TREAT-AD consortium can be part of mutual cross-screening efforts between that group and this. Tim's part of both.