The poster discusses the use of theoretical NMR correlation data in order to see how far NMR correlations could describe a molecule. In this case a molecule is submitted to the WebCocon Server, which will generate theoretical NMR correlation data together with some list-based 13
C chemical shifts (just enough for the rules that Cocon uses). This data is then submitted to Cocon, which should, at least, reproduce the starting molecule. Frequently NMR correlation data will not be enough to fully define a constitution, in this case the suggested alternatives could be used to narrow down how the structure can be better defined. Sometimes just calculating 13
C chemical shifts and comparing them to the experimental values can be enough.
Several examples are shown to illustrate this:
- Caffeine has 6 different constitutions suggested, of which only 2 need to be considered seriously. But these 2 can not be distinguished by calculating 13C chemical shifts;
- Aspirine has 2 different constitutions suggested, one ether (with free acid) and one ester. A possible distinction can be done by derivatization;
- Adrenaline and Ascorbic acid both have no alternative constitutions suggested;
- Maitotoxin has 15 different constitutions suggested. This example was chosen to show that the molecules size is more a problem for the experimental that for the theoretical part.