Error in SMILES produced by Ketcher

[avrilcoghlan]

Hello, I drew a compound into Ketcher on the ChEBI website and saved it as 'Daylight SMILES'.

The SMILES Ketcher gave was C12C=C(N+=O)C(C[C@@]3COC(=O)[C@]3CC3C=C4C(OCO4)=CC=3N+=O)=CC=1OCO2

I've attached a screenshot of the compound I drew in (this is from the original paper about the compound).

However, when I look at this SMILES in CDK depict, it shows that Ketcher has given an incorrect SMILES, as the picture isn't drawn correctly in CDK depict. The problem is that the SMILES is missing hydrogens that should be there. It looks to me like there is a bug in Ketcher that makes it save the SMILES incorrectly.

I then tried drawing the compound in Marvin JS on the Chembl website, and saved the picture as SMILES and it gave me the SMILES O=C1OCC@@H[C@@H]1Cc1cc2c(cc1N+[O-])OCO2

This does seem to be the correct SMILES, as it is drawn correctly when I paste that SMILES in CDK depict.

So it seems to me that Ketcher is producing incorrect SMILES. So it might be better if ChEBI changed to using Marvin JS?

Kind Regards, Avril

[ANiknejad]

Dear ChEBI Team, (@amalik01 (?))

I am wondering if you checked the issue reported by Avril here, because I would like to understand if possibly that could be related to the issues I created recently (issues number 4232 to 4234) talking about missing 2 hydrogens in some formula. And actually if I use the chebi submission tool to check formula generated by SMILES copied from some online chebi compounds, I got a formula different from the one in chebi.

Here an example to hopefully make things clear:

https://www.ebi.ac.uk/chebi/searchId.do?chebiId=CHEBI:77330

if I copy/paste its SMILES in the chebi submission tool (https://www.ebi.ac.uk/chebi/submission/home) the formula appears as C22H29N7O17P3SR But doing the same with the chebi mol file, the formula appears as in CHEBI:77330, so C22H31N7O17P3SR

Many thanks in advance for letting me know your thoughts, Best regards,

Anne

[amalik01]

I have just contacted the team at Ketcher to see if this can be fixed (waiting to hear from them). We use Ketcher since its an open source package unlike Marvin. We currently do not have a dedicated full time developer so at present we will not be able to change to a different structure package since ChEBI is currently being supported on a best effort basis by other members of the team. @ANiknejad In the meantime, please try and avoid submitting structures using SMILES strings since you will end up with the wrong formula.

[amalik01]

Hi, it seems that the latest ketcher version generates the correct SMILES. ChEBI is currently using an older version which needs to be updated. However, we are planning to start the redevelopment of the ChEBI database in the next few months and hopefully will update to the new version as part of this process.