Hi Geoff and Ben, Thanks for the comments! Yes, I fully agree now that my problem is about bond perception, not SMARTS -- I was just describing the symptoms which happened during SMARTS matching.
I would second Ben's suggestion on making ConnectTheDots() more versatile. I understand that having a good default for sanitation is important to make sure that the parsing of PDB and other formats can be done robustly, but unconventional bonding is, as Geoff commented, so common in chemistry that it's probably worthwhile of being considered in ConnectTheDots() via an optional flag. In any case, thanks for all the help! Best regards, Peng On 10/25/2017 11:35 AM, Benjamin Bucior wrote:
I've been working on another system where oxygen frequently exceeds a standard valency of two (organometallics). So far, I've been getting around the automatic bond deletion in an older version of Open Babel by modifying the element.txt file. By changing the MaxBnd column for oxygen to a higher value (e.g. 6), the oxygen atom retains the extra bonds. But that approach obviously causes other problems, including representation of hydrogens. Would it be helpful to have an option in ConnectTheDots to manually disable the max valence cleanup step? For example, rdkit <http://www.rdkit.org/docs/RDKit_Book.html#molecular-sanitization> has flags to selectively disable certain aspects of molecular sanitation. Another question: if you manually build a hydronium ion bond-by-bond, what's the proper way to set the charge for SMILES export? OBAtom::SetFormalCharge? Something with SetTotalCharge? Is there a way to do this automatically on the molecule? Thanks! Ben On Wed, Oct 25, 2017 at 9:51 AM, Geoffrey Hutchison <geoff.hutchi...@gmail.com <mailto:geoff.hutchi...@gmail.com>> wrote: > atom by atom, and then call `ConnectTheDots()` to add bonds Not surprisingly, for a general-purpose library, Open Babel has a general-purpose bond perception algorithm. As you note, it will ensure oxygen atoms have a maximum of two bonds. This ensures PDB and other crystallography files with disordered atoms still retain “normal” valency. > How should I then approach this SMARTS matching problem? My system has regular > water molecules and I have only coordinates and atomic numbers, so a priori I > don't know which oxygen atoms belong to H3O+ and which hydrogen atoms are > bridging hydrogen in H5O2+. I would say *that* is your key problem, not SMARTS. You can certainly adapt the ConnectTheDots code - it simply loops through all atoms and checks where the distance between atoms is smaller than the sum of the covalent radii. Those are “bonds.” (OB’s code will sort along the Z axis and adds a few additional checks for performance reasons, but that’s the basics of the algorithm.) You’ll need a special-purpose bond detection algorithm for your needs. You might find larger hydrogen-bonded clusters beyond H5O2, for example. I’ve found systems with unusual bond distances. That’s part of chemistry. -Geoff ------------------------------------------------------------------------------ Check out the vibrant tech community on one of the world's most engaging tech sites, Slashdot.org! http://sdm.link/slashdot _______________________________________________ OpenBabel-discuss mailing list OpenBabel-discuss@lists.sourceforge.net <mailto:OpenBabel-discuss@lists.sourceforge.net> https://lists.sourceforge.net/lists/listinfo/openbabel-discuss <https://lists.sourceforge.net/lists/listinfo/openbabel-discuss>
------------------------------------------------------------------------------ Check out the vibrant tech community on one of the world's most engaging tech sites, Slashdot.org! http://sdm.link/slashdot _______________________________________________ OpenBabel-discuss mailing list OpenBabel-discuss@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/openbabel-discuss
