PRACTICAL COMPUTER-ASSISTED STRUCTURE ELUCIDATION FOR COMPLEX NATURAL-PRODUCTS - EFFICIENT USE OF AMBIGUOUS 2D NMR CORRELATION INFORMATION

If a computer-assisted primary structure elucidation expert system is to tackle structural problems of real-world complexity, it must be able to cope with the ubiquitous ambiguous 2D NMR correlation information, especially the ambiguous-node(signal or atom)-involved (topological) distance constraints (ANDCs). Such ambiguities of the distance constraints result typically from spectral imperfections and molecular symmetry. As a significant improvement of CISOC-SES, an expert system that automatically deduces the constitutional structure for unknown organic compounds from their molecular formula and conventional one- and two-dimensional NMR spectral information, this paper reports our novel approaches to the representation and use of such ambiguous distance constraints for this purpose. The high performance of the improved CISOC-SES is demonstrated by the structure elucidation of alborixin from its molecular formula (C48H84O14) and H-1, C-13, COSY, HMQC, HMBC, and NOESY spectral data.