Automated Structure Elucidation from 1D C-13 Spectral Data

Antony Williams, Kirill Blinov, Mikhail Elyashberg and Eduard Martirosian

First SMASH NMR Meeting, Chicago, August 1999

Abstract

There have been a number of approaches to structure elucidation using NMR spectroscopy as the foundation spectroscopic technique. Recent approaches have generally used crosspeak responses from multi-dimensional NMR experiments together with classical structure generation from a given molecular formula. We report our recent approach for structure elucidation that can be initiated simply by using a basic 1D spectrum as the input set.

Our Structure Elucidator identifies the structure of an organic compound based on a 1D ¹³C NMR spectrum. Suggested structures are generated from fragment overlap using the unique fragment-based rules system which is part of the our NMR prediction packages which have been successfully applied to the predictions of ¹H, ¹³C, ¹⁹F and ³¹P spectra. Since any particular fragment has associated with it a set of ¹³C chemical shifts, and any spectrum is simply a superposition of spectral subsets with associated fragments, theoretically it is possible to extract suggested fragments with associated chemical shifts and build the full spectrum from appropriate overlap of these fragments. Using additional data including multiplicity data, quantitative intensities, H1 NMR spectral data, infra-red spectral data, mass spectroscopic data and the molecular formula the probability of extracting appropriate fragments and generating the final structure is greatly increased.

During the structure elucidation process the user can participate directly in the process of structure elucidation. The software will display a number of possible structures, with comparison of on-screen experimental and fragment spectra, or, in the case of failure, a set of structural fragments corresponding to portions of the spectrum which can then be used to help assemble the structure of the unknown compound. Using lists of included fragments and excluded fragments, assuming knowledge of the basic synthetic starting materials or partial identification of functional groups and fragments from the spectral data, the structure elucidation maybe biased and simplified.

The Structure Elucidator software also includes a self-training system: if the accuracy of spectral calculations for a new class of compounds is poor, then a user data base can be populated experimental chemical shifts to provide a fragment training set. Subsequently the data base can be used to make calculations for any new and related compound with the program automatically using the information from both the user data base and the internal data base. This power alone enables this particular toolset to be focused on structure elucidation within particular structure classes, especially for those that are not yet reported within the public domain, a common occurrence in todays world of pharmaceutical and chemical research secrecy.

The structure elucidator tool offers the opportunity to process and manipulate NMR, IR and MS data directly from spectrometers using direct vendor formats. Using the processed spectra as the foundation for the structure elucidation the analysis itself can be performed without the provision of any structural information (known substructures or fragments). Under these conditions full elucidation can take from a few minutes to a number of hours depending on the complexity of the task. Using additional constraints such as known fragments and molecular formula elucidation can be performed much quicker.

Download full article in Adobe Acrobat format (1160 Kb PDF file).

Note: To read or print this article you need to have Acrobat Reader. Download the latest version of the Adobe Acrobat Reader.