Hi Christos,The differences are as far as I can see all in the /b layer of the InChI (/b = stereo parity of double bonds), so my guess is that differing 2D coordinates is the cause. Do you also see the difference if you run the output SDF back through KNIME (that is: Are the coordinates generated by SDF production different than earlier conformations used to calculate InChI ?) ?
I don't know if subtle RDKit version differences between KNIME and conda could cause it - I am not familiar enough with the KNIME nodes to be able to judge that.
Cheers -- Jan Holst Jensen On 2019-11-12 13:14, Christos Kannas wrote:
Dear RDKiters, I'm having the following problem.I have a workflow that standardises compounds and as part of the process it generates standard InChI and InChIkey for the compound. The output is stored in an SDF. If I parse the SDF to a dataframe in jupyter notebook, then use the mol object to generate standard inchi, for a small number of compounds the new standard InChI is slightly different than the one generated in Knime environment.Environments Details: - RDKit Knime Nodes: 3.8.0v201906261723 - RDKit Python (conda): 2018.09.3, 2019.03.4, 2019.09.1 See image: https://imgur.com/a/EnYoHWG Best, Christos Christos Kannas Scientific Software Developer (Cheminformatics)
smime.p7s
Description: S/MIME Cryptographic Signature
_______________________________________________ Rdkit-discuss mailing list Rdkit-discuss@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/rdkit-discuss
