Effects of the application of different window functions and projection methods on processing of 1H J-resolved nuclear magnetic resonance spectra for metabolomics.

Two dimensional (2D) homonuclear (1)H J-resolved (JRES) nuclear magnetic resonance spectroscopy is increasingly used in metabolomics. This approach visualises metabolite chemical shifts and scalar couplings along different spectral dimensions, thereby increasing peak dispersion and facilitating spectral assignments and accurate quantification. Here, we optimise the processing of 2D JRES spectra by evaluating different window functions, a traditional sine-bell (SINE) and a combined sine-bell-exponential (SEM) function. Furthermore, we evaluate different projection methods for generating 1D projected spectra (pJRES). Spectra were recorded from three disparate types of biological samples and evaluated in terms of sensitivity, reproducibility and resolution. Overall, the SEM window function yielded considerably higher sensitivity and comparable spectral reproducibility and resolution compared to SINE, for both 1D pJRES and 2D JRES datasets. Furthermore, for pJRES spectra, the highest spectral quality was obtained using SEM combined with skyline projection. These improvements lend further support to utilising 2D J-resolved spectroscopy in metabolomics.