Questions on assay results csv

I’ve been looking at the assay results at and have a few questions.

Is there an explanation for the column names in the csv file? Most of these make sense, but I’m not sure what the difference is between ‘IC50 (uM)’ and ‘IC50 (uM) - std’.

What were the criteria for deciding which compounds were tested in dose-response?

Several of the structures are duplicated. They have the same SMILES, but different CIDs
CC(=O)N1CCN(CC(=O)Nc2cccc(S(N)(=O)=O)c2)CC1 JAN-GHE-1d98ec1c-7,LON-WEI-b8d98729-4
CC(C(=O)O)c1ccc2c(c1)[nH]c1ccc(Cl)cc12 AAR-POS-fca48359-19,AAR-POS-fca48359-20,GER-UNI-ec786817-1,MAK-UNK-0d6072ac-4
CC©N©C(=O)C1CCN(C(=O)CCl)CC1 AAR-POS-0daf6b7e-19,VIR-GIT-7b3d3065-3
CNS(=O)(=O)Cc1ccc2[nH]cc(CCN©C)c2c1 ANT-OPE-47f3bb65-1,TOM-OIS-cbf37265-2,LON-WEI-b8d98729-34
COc1ccc2[nH]cc(CCNC©=O)c2c1 ALE-ACE-2610ee4d-1,JOH-MSK-35a8745a-1
CS(=O)(=O)Nc1cccc(C(=O)Nc2cccc(N3CCCC3=O)c2)c1 MAK-UNK-1e8f9e3c-1,MAK-UNK-9159f8ed-6
Cc1ccc©c(S(=O)(=O)N2CCN(C(=O)CCl)CC2)c1 AAR-POS-0daf6b7e-3,WAR-XCH-79d12f6e-6
Cc1ccncc1NC(=O)Nc1cccc(Cl)c1 JAG-UCB-cedd89ab-8,TRY-UNI-714a760b-12
Cc1ccncc1NC(=O)Nc1ccccc1 TRY-UNI-714a760b-4,ALE-HEI-f28a35b5-1
Cc1ccncc1NC(=O)c1cccc(NC(=O)Nc2cccnc2)c1 MAK-UNK-129dcd6f-9,MAK-UNK-129dcd6f-11
N#Cc1ccc(NC(=O)Nc2cccnc2)cc1 JOH-UNK-14e6adc5-2,ANN-UNI-26382800-3
N#Cc1cccc(NC(=O)Nc2cccnc2)c1 JOR-UNI-61e7b1d5-2,ANN-UNI-26382800-4,WAR-XCH-eb7b662f-1,DAR-DIA-23aa0b97-5
O=C(CCl)N1CCC(C(=O)N2CCCCC2)CC1 AAR-POS-0daf6b7e-9,VIR-GIT-7b3d3065-2
O=C(CCl)N1CCN(Cc2cccc(Cl)c2)CC1 AAR-POS-d2a4d1df-35,MED-COV-4280ac29-15,MED-COV-4280ac29-16
O=C(CCl)N1CCN(Cc2cccc3ccccc23)CC1 AAR-POS-d2a4d1df-40,MED-COV-4280ac29-31
O=C(CCl)N1CCN(Cc2cccs2)CC1 AAR-POS-0daf6b7e-17,LON-WEI-8f408cad-5
O=C(CCl)N1CCN(Cc2ccsc2)CC1 MAK-UNK-212f693e-8,AAR-POS-0daf6b7e-13,LON-WEI-8f408cad-7
O=C(CCl)N1CCN(S(=O)(=O)c2c(F)cccc2F)CC1 AAR-POS-d2a4d1df-32,LON-WEI-8f408cad-2
O=C(CCl)N1CCN(S(=O)(=O)c2cccc(F)c2)CC1 AAR-POS-d2a4d1df-23,LON-WEI-8f408cad-4
O=C(CCl)N1CCN(S(=O)(=O)c2cccs2)CC1 AAR-POS-d2a4d1df-22,LON-WEI-8f408cad-3
O=C(CCl)Nc1cccc(N2CCCC2=O)c1 MAK-UNK-7a704a63-10,AAR-POS-0daf6b7e-11,LON-WEI-8f408cad-1
O=C(c1cc(=O)[nH]c2ccccc12)N1CCN(c2ccccc2)CC1 EDJ-MED-78f964c8-1,MAT-POS-916a2c5a-4

Any idea what’s going on here?



Hi @pwalters, good questions!

The explanation is bad naming. This is being fixed (‘IC50 (uM) - std’ will soon read ‘IC50 (uM) - StdDev’ ). We are also in the process of fixing an error with these standard deviations, which are currently incorrect. The value in that sheet is currently ~1.96*\sigma. Note that we also just fixed a column mapping error, so you may want to re-export. Sorry!

If you want the standard deviations right now, the plots at least have the correct confidence intervals right now.

Generally, if they exhibit above 50% inhibition at 50 uM, we have continued on to do the whole curve. We will probably make this cutoff more strict to above 50% inhibition at 20 uM in future rounds.

This is unfortunately an intricacy of the way we setup the submissions process. We allow multiple users to submit the same molecule design idea. This has some advantages, such as knowing when a particular design has been chosen by multiple chemists. However, the main disadvantage is that there can be multiple CIDs associated with a particular molecule. Therefore, we have the canonical_CID listed, which is the first ID associated with that molecule. Thus, CIDs that share a molecule also share a canonical_CID. For example,
JOR-UNI-61e7b1d5-2,ANN-UNI-26382800-4,WAR-XCH-eb7b662f-1,DAR-DIA-23aa0b97-5 all share the canonical_CID of DAR-DIA-23aa0b97-5

If we were rapidly setting up a crowdsourcing initiative a second time, I might have a few changes - but I guess we are where we are.

Sorry again for the confusion and the slight error, I hope to have this fixed soon and please let me know if I can help in any other way.


1 Like

Thank you Matt for your rapid and complete answer to my questions!


1 Like