> I've been trying to use Open Babel (windows) to perform a file conversion > from a log file (Gaussian) to a Tinker xyz file, while demanding the > conversion to work with MM3 atom types. The problem began with trying to > convert an ordinary primary amide (expecting atom type 9 in return) log file, > which kept giving the atom type 151 in return. My intention is to build an > automated sequence of calculations which relies heavily on your magnificent > tool, and sadly I've been having the same difficulty with any amide structure > (essentially my entire data set is made of amides).
This isn't really "information loss." Consider that the Gaussian log file has no bonding information - it's a collection of atoms. so Open Babel has to perceive bond types before it can determine atom types. So the first question I have is "did Open Babel find the correct bond orders in your structure?" For example, open up the Gaussian log file in Avogadro (which uses Open Babel) and make sure the bond orders look good. Or use OB to convert the Gaussian output to a Mol2 or SDF and open in a different program to make sure the bonds have the right orders (e.g, there's a C=O). But I'd be quite certain that the MM3 atom types aren't always assigned correctly. Very few people use MM3 and so I doubt it's gotten very thorough debugging. If you can give me some example files with corrected atom types, it's a pretty easy thing to fix. But first, make sure the bonding is perceived correctly. -Geoff _______________________________________________ OpenBabel-discuss mailing list OpenBabel-discuss@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/openbabel-discuss
