PostEra

Instructions For Selecting Non Covalent Ideas For Synthesis By Wednesday 20th May

AIM:
Select a total of around 500 non covalent compounds for synthesis based primarily on ranking from the computational method used to model the crowd sourced ideas.

DEADLINE:
Compound selections from each method must be made by Wednesday 20th May at the very latest and the earlier a selection can be submitted the better.

The process for selecting compounds is as follows:

  • Each person who has run a computational calculation on the set of crowd sourced, non covalent compounds will rank everything that has been modelled using the results of their calculation. Ideally everyone has modelled the full set of crowd sourced ideas (~4000 compounds), but if you have modelled less then rank as many as you have modelled using your chosen scoring function/methodology.

  • Post an sdf of your ranked list of compounds in Fragalysis format to your forum post describing your method.

  • If possible include an sdf with the full ranked list and a separate sdf containing your top ranked selected compounds aiming for a selection of around 100-200 compounds.

  • Your selection will be passed on to the medicinal chemistry team.

  • Your selection of compounds should not be biased towards personal preferences, selection should be primarily made using your chosen computational scoring method.

  • Each list of compounds from each methodology will be eyeballed by the medicinal chemists for a sanity check, compounds will only be removed if something looks obviously wrong with a compound, again no personal preferences or bias will be used in removing compounds.

  • Compound lists will be run through a synthetic accessibility filter.

  • Equal numbers of compounds will then be selected from each computational method based on their rank to make up the final selection for synthesis.

  • If med chem eyeballing and synthetic accessibility reduces a methods number of compounds below the required number from any particular computational method, the list can be topped up with the next highest ranked compounds from that particular method which pass med chem eyeballing and synthetic accessibility.

  • The final combined selection of around 500 compounds will then be submitted for synthesis at a CRO.

Once synthesised and tested we will post analyse the results to determine which computational method has been successful in identifying hit molecules. This will guide us towards which is the best methodology to apply in modelling this protein in the future.

4 Likes

I have also searched for commercially available compounds which contain fragments (such as NC=1C=NC=NC1) and docked these using the THINK software which was used for the Oxford University cancer and Find-a-Drug distributed computing projects about 20 years ago (https://treweren.com). Would these of of interest at this time?

I’m about 90% through docking the submitted molecules. I had planned to cluster these based on volume overlap but I doubt whether I’ll get that done by Wednesday so I’ll aim to rank them and submit them.

Thanks for this! If I’ve understood correctly, the compounds you ranked are not necessarily Moonshot compounds (ones already submitted) but commercially available ones containing relevant fragments? If so, can I ask how many you have as it’s important they are submitted through through the Moonshot platform first.

Thanks.

The short answer is that there are thousands of predicted hits containing C=1C=NC=NC1 (which is x0995 without the chain N) and hundreds of hits containing O=CNC=2C=CC=NC2 (which is x0678 without the methylcyclohexane). I also have a couple of hundred hits containing CCC=1C=CN=CC1 (which is the ethylpyridine from x0540). I havent considered looking at these by eye or for sources other than ChemBridge.

I’m currently attempting to hack my ranked SD file for the submitted molecules into the format and with the fields you want. When this is complete I will take a closer look at the hits from ChemBridge and submit at least the top 100 of these.

I’m checking pricing etc.

Thanks for your post. Would it be possible for you to take a look at this post on the forum: Instructions around the Docking Results category

It gives instructions for how to correctly format and upload your data for it to be more easily included in compound selection.

Many thanks.

I’ve made some preliminary enquiries about purchase pricing for compounds from available collections. Basically for 100 it’s about $20 per compound per microMole reducing to about half of that for high numbers with exceptions for low stock compounds which are at a higher price (VAT, Duty and shipping are extra). Some of compounds (I know which) are currently in DMSO which would incur higher costs for either shipping or drying!

Okay, thanks for the pricing info and for your docking results post (which I have just seen)!