Curve-fitting of Fourier manipulated spectra comprising apodization, smoothing, derivation and deconvolution

We present a general method for curve-fitting Fourier manipulated spectra, comprising apodized, smoothed, derivatised and deconvoluted spectra. The analytical expressions of Fourier manipulated bands in the spectral domain, needed for the curve-fitting, are usually very complex or do not even exist; hence an accurate curve-fit of Fourier manipulated spectra becomes unfeasible. Our strategy is to construct both the model and their derivatives in the Fourier domain, where they have simple and general expressions, and then Fourier transform them back to the spectral domain. The first benefit of this approach is the accurate curve-fitting of Fourier deconvoluted spectra, a main step in the secondary structure estimation of proteins by FTIR spectroscopy. © 2004 Elsevier B.V. All rights reserved.