13C NMR Spectral Prediction by Means of Generalized Atom Center Fragment Method

Knowledge-based NMR spectral prediction relies on the correlations between substructures and sub-spectra. To extract the correlations, a systematic substructure measurement has been developed to classify substructures according to their chemical shift values. Historically, the atom center fragment (ACF) concept has been used as a means to systematically measure substructures for NMR spectral prediction. The assumption behind this concept is that the chemical shift value of an atom is influenced by its chemical environment. Based upon the study of the ACF-type approaches, a generalized atom center fragment (GACF) approach is proposed in this paper. In the GACF approach, a substructure consists of a center atom, core layer, and external layers. The center atom and the core layer, are identified as the super center atom. The external layers are the chemical environment. A number of algorithms have been developed to measure GACF substructures from a structure database, and create the NMR knowledge base for NMR spectral prediction.