Maximum likelihood constrained deconvolution. I: Algorithm and qualitative and quantitative enhancement in synthetic two‐dimensional NMR spectra

Abstract

AbstractThe maximum likelihood method is a constrained iterative spectral deconvolution technique in which a spectral fitting model is determined by minimizing the variance of fit in the time domain in a nonlinear iterative manner. Application of this method to synthetic 2‐dimensional (2‐D) NMR spectra, which have heavily overlapped multiplets associated with low signal to noise ratios, yields contrast‐enhanced spectra with simultaneous noise suppression and resolution improvement. This protocol greatly facilitates peak recognition and often partitions overlapping multiplets into individual components, leading to a more accurate interpretation of resonance frequencies, coupling constants, and multiplets than does the conventional apodization or Fourier transform method. These advantages are useful for constructing reliable 3‐D molecular structures for complex molecular systems. © 2002 Wiley Periodicals, Inc. Concepts Magn Reson 14: 402–415, 2002