Blind separation of analytes in nuclear magnetic resonance spectroscopy: Improved model for nonnegative matrix factorization

Ivica Kopriva, Ivanka Jerić
2014 Chemometrics and Intelligent Laboratory Systems  
We introduce improved model for sparseness constrained nonnegative matrix factorization (sNMF) of amplitude mixtures nuclear magnetic resonance (NMR) spectra into greater number of component spectra. In proposed method selected sNMF algorithm is applied to the square of the amplitude of the mixtures NMR spectra instead to the amplitude spectra itself. Afterwards, the square roots of separated squares of components spectra and concentration matrix yield estimates of the true components amplitude
more » ... spectra and of concentration matrix. Proposed model 2 remains linear in average when number of overlapping components is increasing, while model based on amplitude spectra of the mixtures moves away from the linear one when number of overlapping components is increased. That is demonstrated through conducted sensitivity analysis. Thus, proposed model improves capability of the sparse NMF algorithms to separate correlated (overlapping) components spectra from smaller number of mixtures NMR spectra. That is demonstrated on two experimental scenarios: extraction of three correlated components spectra from two 1 H NMR mixtures spectra and extraction of four correlated components spectra from three COSY NMR mixtures spectra. Proposed method can increase efficiency in spectral library search by reducing occurrence of false positives and false negatives. That, in turn, can yield better accuracy in biomarker identification studies which makes proposed method important for natural products research and the field of metabolic studies.
doi:10.1016/j.chemolab.2014.06.004 fatcat:dlieb5m6wrcptemudce7act7ny