Standardizer Stripping Stereochemistry

User 535731474a

22-08-2014 16:37:05

Hi,

It appears that standardizer is stripping stereochemistry from CHEMBL100.

Attached is the standardizer config we used and the original CHEMBL100 in sdf format ("chembl100_original.sdf")

chemaxon/jchem/bin/standardize -c standardizer.xml -f smiles:un -o cheml100.smi chembl100.sdf

The output is the following without stereochemistry information

CC1(C)Oc2ccc(cc2C(C1O)N1CCCC1=O)C#N

It should output

CC1(C)Oc2ccc(cc2[C@H]([C@@H]1O)N1CCCC1=O)C#N

We have attached a working sdf that does render the correct smiles after standardization.( chembl100_working.sdf)

Is there something wrong with the chembl mol format?

ChemAxon d26931946c

25-08-2014 12:52:36

Hi Scott,

I think you've mixed up the original and the working files.

The problem is in the header line of the file:

SciTegic01161211223D

This tells our importer to interpret the file as a 3D molecule, but the coordinates are missing for the third dimension, therefore the parity information is missing.

If you change this line to

SciTegic01161211222D

then the stereo information is preserved.

I could not find this sdf file in ChemBl, all the downloaded files had the correct header line. Could you let us know how you've downloaded this file?

Best regards,

Peter