Using maximum likelihood spectral deconvolution in multidimensional nuclear magnetic resonance

This chapter describes the maximum likelihood method (MLM) at a practitioner's level, illustrates results of MLM spectral deconvolution under conditions typical of two-dimensional (2D) and three-dimensional (3D) NMR experiments on nucleic acids and proteins, and offers advice on optimal application of the method. Quantification of frequencies and intensities is excellent for cross-peaks determined by MLM and a related protocol, symmetrized maximum likelihood (SML) deconvolution, the latter exploiting the symmetry properties of the experimental spectrum together with MLM deconvolution Maximum likelihood reconstruction is easily applied to a wide range of multidimensional NMR spectra. There are few parameters to specify, and the method should find most important use in relatively small, crowded spectral regions. The frequency fidelity of the reconstruction is excellent, as is the quantitative integration of peaks, provided the protocols described are adhered to. Although the most conspicuous feature of the reconstructed spectra is the apparent reduction in noise, the most useful aspect is actually the sharpening of spectral features. The latter is a great aid to assignment and integration of peaks in heavily overlapped regions.