Molecule to CDK node gives missing object with valid SMILES

[gcincilla]

Hi guys,

I have a problem with the “Molecule to CDK” node. As you can see from the attached workflow, there are some molecules, provided through valid SMILES, that cannot be parsed by the node. It gives “Object missing” without any error or warning in the log. Unfortunately I can not understand when this happens and especially why. This problem is not related to this previously solved bug. I'm using the last KNIME-CDK version (i.e. 1.5.600.v201607162102).

Please, could you confirm that it is a bug?

Thanks

Gio

• mol2cdk_problem_2.knwf

[Stephan]

Hi Gio,

I could reproduce the result. The problem appears to be the indole system in both cases. The Nitrogen heteroatom is simply encoded as 'n' rather than '[nH]'. That's the problem.

BrCC(=O)NC1(C)CCC(CC1)C(NCC(O)COc2cccc3c2ccn3)(C)C <p>BrCC(=O)NC1(C)CCC(CC1)C(NCC(O)COc2cccc3c2cc[nH]3)(C)C</p> </td> </tr> </tbody>

Are you certain that the SMILES you provided are valid? I generated a couple of indole SMILES using other tools and always got the '[nH]' variant.

Kind regards,

Stephan

[gcincilla]

You are completely right Stephan,

You are completely right Stephan, the indole must be written with the "[nH]".  I didn't realize that was the problem. This is what happens when you don't double check the SMILES coming from potentially erroneous sources.

Thank you very much for checking this.

Cheers,

Gio