Short-time Fourier Transform of Free Induction Decays for the Analysis of Serum Using Proton Nuclear Magnetic Resonance

Abstract

Serum is the liquid portion of blood routinely used as a clinical specimen. Serum consists of low-molecular-weight compounds 1 3 , macromolecules and their degradation products 4, 5 , inorganic ions 6, 7 , and water. Hydrogen nuclei in serum play important roles through intra-and intermolecular interactions with other nuclei. Most of the hydrogen atoms in serum are associated with water molecules or proteins. Water molecules form hydrogen bonds with nitrogen and hydrogen atoms in macromolecules and low-molecular-weight compounds, stabilizing their molecular conformations 8 11. Hydrogen bonding also takes place in the interactions between solutes. Thus, serum has its own dynamic features and assumes a supramolecular structure

Time–frequency analysis of serum with proton nuclear magnetic resonance for diagnosis of pancreatic cancer

“…Methods for 1 H-NMR measurement and FID analysis using STFT are described elsewhere 28 . Briefly, each serum sample (200 μL) was mixed with 400 μL of deuterium oxide ( 2 H 2 O) (ISOTEC, Sigma-Aldrich, St. Louis, MO, USA) and pipetted into a 5-mm (outer diameter) NMR tube (Wilmad-LabGlass, Vineland, NJ, USA) for NMR analysis.…”

“…Data calculated using STFT were imported into Unscrambler X software (version 10.5; Camo Software AS, Oslo, Norway) for a partial least-squares discriminant analysis (PLS-DA) 28 . All data were transformed using standard normal variates to remove scatter effects, by centering and scaling each individual variable.…”

“…Cross-validation was used to statistically evaluate and compare learning algorithms by randomly dividing the data into a software-specified optimal number of segments. To estimate the goodness of fit for future predictions using a defined number of factors, R 2 and Q 2 were calculated for the explained-variance plots for calibration and validation data sets, respectively 28 .…”

“…Short-time Fourier transform (STFT) is a basic method for time–frequency analyses 24 – 27 . We have recently reported that the STFT of FIDs obtained from proton NMR ( 1 H-NMR) measurements offers a useful serological diagnostic tool 28 . No previous reports appear to have applied this analysis as a diagnostic tool for diseases, including PC.…”

Signal Deconvolution and Noise Factor Analysis Based on a Combination of Time–Frequency Analysis and Probabilistic Sparse Matrix Factorization

“…STFT has the ability to extract time-varying behavior from FIDs, allowing for the analysis of dynamic chemical shifts of atoms in flexible proteins [26]. In addition, it has been reported that STFT can extract T 2 * information from FIDs and improve the results of discriminant analysis [27]. Applying the same idea to covariance NMR [28], T 2 *-weighted covariance NMR improves the sensitivity and resolution of signals based on the difference in T 2 *, determined by dividing each FID in the t 1 dimension of 2D-NMR to create a series of sub-FIDs [29].…”