If you use canonical SMILES instead of the InChI to evaluate changes in stereo, do the results change? If you use canonical SMILES, you will also be able to manually verify what's different.
- Noel On Fri, 21 Jun 2019 at 09:05, Naruki Yoshikawa <naruki.yoshik...@gmail.com> wrote: > Hi all, > > I've been developing a new distance geometry method these days. > https://github.com/openbabel/openbabel/pull/1875 > > I generated SMILES from experimental structures and predict a 3D > structure from the SMILES. > I compare InChIKey of experimental structure and predicted structure > to evaluate stereochemistry correctness. > The distance geometry function uses > OBDistanceGeometry::CheckStereoConstraints() to check stereo > constraints are satisfied. > > https://github.com/openbabel/openbabel/blob/528f73fe64376ce8922897748e2879a6d599e7ed/src/distgeom.cpp#L953 > Even though stereo constraints are satisfied judging from this > function, InChIKey did not match in some molecules. > > I prepared test cases here: > > https://gist.github.com/n-yoshikawa/00dc043694b539fb5f9340f48fb5d2b0#file-checkstereo-cpp > This gist includes SDFs predicted from SMILES and they have the above > problem as shown in result.txt. > > Could you suggest possible cause of this problem? > > Naruki > > > _______________________________________________ > OpenBabel-Devel mailing list > OpenBabel-Devel@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/openbabel-devel >
_______________________________________________ OpenBabel-Devel mailing list OpenBabel-Devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/openbabel-devel
