Maximum-Likelihood-Constrained Deconvolution of Two-Dimensional NMR Spectra. Accuracy of Spectral Quantification

Abstract

The maximum-likelihood method (MLM) and a related processing protocol involving diagonally symmetrized data, symmetrized maximum-likelihood (SML) spectral deconvolution, yield contrast-enhanced 2D NMR spectra with simultaneous noise suppression and marked resolution enhancement, relative to spectra processed with conventional apodization and FFT. In this paper the MLM deconvolution technique and SML protocol were applied to synthetic and real 2D NOE spectra to evaluate the accuracy of peak-volume quantification under various conditions. It is concluded that use of MLM and SML is remarkably robust for peak discrimination and volume calculations in 2D NMR studies, despite the inherent nonlinear nature of MLM. These protocols clearly facilitate identification and quantification of heavily overlapped cross peaks. In this study, quantitative integration linearity was observed at peak dynamic ranges of up to 1000:1 for 2D spectral reconstructions.