hrvatski jezikClear Cookie - decide language by browser settings

Library-Assisted Nonlinear Underdetermined Blind Separation and Annotation of Pure Components from 1H Nuclear Magnetic Resonance Mixture Spectra

Kopriva, Ivica; Jerić, Ivanka; Popović Hadžija, Marijana; Hažija, Mirko (2019) Library-Assisted Nonlinear Underdetermined Blind Separation and Annotation of Pure Components from 1H Nuclear Magnetic Resonance Mixture Spectra. Analytica Chimica Acta, 1080 . pp. 55-65. ISSN 0003-2670

PDF - Accepted Version - article
Download (892kB) | Preview


Due to its capability for high-throughput screening 1H nuclear magnetic resonance (NMR) spectroscopy is commonly used for metabolite research. However, the key problem in 1H NMR spectroscopy of multicomponent mixtures is overlapping of component signals and that is increasing with the number of components, their complexity and structural similarity. This makes metabolic profiling, that is carried out through matching acquired spectra with metabolites from the library, a hard problem. Here, we propose a methodology for nonlinear blind separation of highly correlated components spectra from smaller number of, including one only, 1H NMR mixture spectra. The method transforms related nonnegative underdetermined blind source separation problem into multiple high-dimensional reproducible kernel Hilbert Spaces (mRKHSs). Therein, highly correlated components are separated by sparseness constrained nonnegative matrix factorization (sNMF) in each induced RKHS. Afterwards, analytes are identified through comparison of separated components with the library comprised of 160 pure components, whereas significant number of them is expected to be related with diabetes. The method is exemplified on: (i) annotation of five components spectra separated from two and one 1H NMR model mixture spectra; (ii) annotation of 55 metabolites separated from 1H NMR mixture spectra of urine of subjects with and without diabetes type 2. Arguably, it is for the first time to propose method for blind separation of large number of components from single nonlinear mixture. Moreover, proposed method pinpoints urinary creatine, glutamic acid and 5-hydroxyindoleacetic acid as the most prominent metabolites in samples from diabetic subjects, when compared to healthy controls. We also provide metabolic interpretation of obtained results.

Item Type: Article
Additional Information: The work performed has been supported through grant IP-2016-06-5235 "Structured decompositions of empirical data for computationally-assisted diagnosis of disease" funded by the Croatian Science Foundation. The first author thanks Gary McGuire for language editing of the final version of the manuscript.
Uncontrolled Keywords: nonlinear underdetermined blind source separation; single mixture; multiple reproducible kernel Hilbert spaces; nonnegative sparse matrix factorization; 1H NMR spectroscopy; metabolic profiling
Subjects: NATURAL SCIENCES > Mathematics > Numerical Mathematics
Divisions: Division of Electronics
Project titleProject leaderProject codeProject type
Strukturne dekompozicije empirijskih podataka za računalno potpomognutu dijagnostiku bolesti-DEDADIvica KoprivaIP-2016-06-5235HRZZ
Depositing User: Ivica Kopriva
Date Deposited: 26 Feb 2020 11:48
DOI: 10.1016/j.aca.2019.07.004

Actions (login required)

View Item View Item


Downloads per month over past year

Increase Font
Decrease Font
Dyslexic Font