Acquiring routine 2D NMRe spectra is a time-consuming process and does not always lead to the desired results. Hyphenated 2D NMR data then come into play, but usually at a late stage in the proceedings because they are not always needed, are time-consuming to acquire and, in some cases, simply because scientists are not yet comfortable in interpreting hyphenated data. NMR spectroscopists face a dilemma when it comes to obtaining the additional information sooner or later - should they or shouldn't they? The hyphenated techniques add considerably to the workload and may be superfluous, but deferring the decision until later means additional work once the initial data reveal themselves to be insufficient.

What would it mean to a spectroscopist if they didn't have to acquire additional data? What if they could simply algorithmically combine the data they had already acquired to generate a data set that is not only equivalent in data content to that requiring hours of acquisition but in many case superior? Collaborative work between NMR expert Gary Martin of Schering-Plough in New Jersey and ACD/Labs has enabled this approach.

"When it is realized that the hyphenated data would be useful, it is usually too late to return to the starting point to re-run experiments. Usually, at that point, you have several potential choices", explains Martin. "The first is to locate the sample and acquire those hyphenated data, assuming that the sample is still intact and has not decomposed. The second option is to isolate more material and acquire the data if no original sample is available. But, there is a third option" he says, " which involves calculation as opposed to experimentation."

"Many times routine 2D NMR spectral data are accumulated but more time consuming, hyphenated 2D NMR data are not acquired to minimize spectrometer time/project," Martin told us, "Quite often an investigator will move on to begin acquiring other data and may not come back to the acquired data for several days or longer."

Now, Martin and his colleagues have developed a novel approach to NMR data gathering that helps reduce the workload without invoking Sod's Law. They have found that even at the late stage, following basic 2D acquisition, they can apply Unsymmetrical Indirect Covariance processing methods to simply calculate hyphenated spectra that would take many hours to acquire in just a few seconds. So spectral pairs such as GHSQC and COSY spectra can be combined to generate a data set equivalent in information content to a GHSQC-COSY spectra. The decision to carry out the actual GHSQC-COSY can then be made only if there are irreconcilable overlap problems with the virtual spectra and the 2D results.

"Using Unsymmetrical Indirect Covariance processing methods allows us to calculate a GHSQC-COSY spectrum from already acquired GHSQC and COSY spectra," explains Martin, "the latter will almost always be available, which avoids having to resort to acquiring the actual hyphenated 2D NMR data set unless there are problems due to resonance overlap with the calculated spectrum." He adds that this approach represents a rather considerable time saving, in fact many orders of magnitude. In some examples Martin has generated an equivalent dataset in less than 5 seconds that would usually take an overnight run to acquire.

The calculated GHSQC-COSY spectra can afford signal to noise ratios ranging from about 20 to 160 times higher than the acquired data in the teams test runs. The same UICP approach also provides the means for determining long-range ¹³C-¹³C connectivity networks, adds Martin. These long-range connectivity data are usually the preserve of m,n-ADEQUATE experiments. Moreover, they can also calculate ¹³C-¹⁵N correlation information, which have no experimental equivalent given the statistical improbability of finding both a ¹³C and a ¹⁵N in the same molecular skeleton.

"We can make available data for small molecules that have never before been reported at natural abundances," he adds. "While we have advanced this work significantly in the past months this work is really in its infancy. In terms of a contribution to NMR science this emphasizes the power of software and algorithmic manipulation to the value of extracting more value out of already existing data".

Related links:

• Magn. Reson. Chem in press
• Martin Company Site
• ACD/Labs NMR Prediction 
• Interview with Martin 

Article by David Bradley