Geographical discrimination of Asian red pepper powders using 1H NMR spectroscopy and deep learning-based convolution neural networks.

This study investigated an innovative approach to discriminate the geographical origins of Asian red pepper powders by analyzing one-dimensional 1H NMR spectra through a deep learning-based convolution neural network (CNN). 1H NMR spectra were collected from 300 samples originating from China, Korea, and Vietnam and used as input data. Principal component analysis - linear discriminant analysis and support vector machine models were employed for comparison. Bayesian optimization was used for hyperparameter optimization, and cross-validation was performed to prevent overfitting. As a result, all three models discriminated the origins of the test samples with over 95 % accuracy. Specifically, the CNN models achieved a 100 % accuracy rate. Gradient-weighted class activation mapping analysis verified that the CNN models recognized the origins of the samples based on variations in metabolite distributions. This research demonstrated the potential of deep learning-based classification of 1H NMR spectra as an accurate and reliable approach for determining the geographical origins of various foods.

Development and Validation of a Simple Method to Quantify Contents of Phospholipids in Krill Oil by Fourier-Transform Infrared Spectroscopy.

This study focuses on developing a quantification method for phosphatidylcholine (PC) and total phospholipid (PL) in krill oil using Fourier-transform infrared (FT-IR) spectroscopy. Signals derived from the choline and phosphate groups were selected as indicator variables for determining PC and total PL content; calibration curves with a correlation coefficient of >0.988 were constructed with calibration samples prepared by mixing krill oil raw material and fish oil in different ratios. The limit of detection (LOD, 0.35-3.29%) of the method was suitable for the designed assay with good accuracy (97.90-100.33%). The relative standard deviations for repeatability (0.90-2.31%) were acceptable. Therefore, both the methods using absorbance and that using second-derivative were confirmed to be suitable for quantitative analysis. When applying this method to test samples, including supplements, the PC content and total PL content were in good agreement with an average difference of 2-3% compared to the 31P NMR method. These results confirmed that the FT-IR method can be used as a convenient and rapid alternative to the 31P NMR method for quantifying PLs in krill oil.

A Second Derivative Fourier-Transform Infrared Spectroscopy Method to Discriminate Perilla Oil Authenticity.

The aim of this study was to discriminate the authenticity of perilla oils distributed in Korea using their Fourier-Transform infrared spectroscopy (FT-IR) spectra with attenuated total reflectance accessory. By using orthogonal projections for latent structures discriminant analysis (OPLS-DA) technique, the =C-H cis-double bond, -C-H asymmetric and -C-H symmetric stretching are determined to be the best variables for discriminating the perilla oil authenticity. Comparing the integral and the second derivative methods between authentic and adulterated perilla oil samples, the most obvious and significant differences among the three variables is =C-H cis-double bond stretching. The procedure for applying the second derivative range of variables found in authentic perilla oil samples correctly discriminated between the adulterated samples of perilla oils with soybean oils and/or corn oils added at concentrations of ≥ 5 vol%. These results showed that the second derivative FT-IR analysis can be used as a simple and alternative method for discriminating the authenticity of perilla oil.