Two-dimensional visible/near-infrared correlation spectroscopy study of thermal treatment of chicken meats

Robust variable selection in the framework of classification with label noise and outliers: Applications to spectroscopic data in agri-food

Classification of high-dimensional spectroscopic data is a common task in analytical chemistry. Well-established procedures like support vector machines (SVMs) and partial least squares discriminant analysis (PLS-DA) are the most common methods for tackling this supervised learning problem. Nonetheless, interpretation of these models remains sometimes difficult, and solutions based on feature selection are often adopted as they lead to the automatic identification of the most informative wavelengths. Unfortunately, for some delicate applications like food authenticity, mislabeled and adulterated spectra occur both in the calibration and/or validation sets, with dramatic effects on the model development, its prediction accuracy and robustness. Motivated by these issues, the present paper proposes a robust model-based method that simultaneously performs variable selection, outliers and label noise detection. We demonstrate the effectiveness of our proposal in dealing with three agri-food spectroscopic studies, where several forms of perturbations are considered. Our approach succeeds in diminishing problem complexity, identifying anomalous spectra and attaining competitive predictive accuracy considering a very low number of selected wavelengths.

Rapid Classification of Wheat Grain Varieties Using Hyperspectral Imaging and Chemometrics

The classification of wheat grain varieties is of great value because its high purity is the yield and quality guarantee. In this study, hyperspectral imaging combined with the chemometric methods was applied to explore and implement the varieties classification of wheat seeds. The hyperspectral images of all the samples covering 874–1734 nm bands were collected. Exploratory analysis was first carried out while using principal component analysis (PCA) and linear discrimination analysis (LDA). Spectral preprocessing methods including standard normal variate (SNV), multiplicative scatter correction (MSC), and wavelet transform (WT) were introduced, and their effects on discriminant models were studied to eliminate the interference of instrumental and environmental factors. PCA loading, successive projections algorithm (SPA), and random frog (RF) were applied to extract feature wavelengths for redundancy elimination owing to the possibility of existing redundant spectral information. Classification models were developed based on full wavelengths and feature wavelengths using LDA, support vector machine (SVM), and extreme learning machine (ELM). This optimal model was finally utilized to generate visualization map to observe the classification performance intuitively. When comparing with other models, ELM based on full wavelengths achieved the best accuracy up to 91.3%. The overall results suggested that hyperspectral imaging was a potential tool for the rapid and accurate identification of wheat varieties, which could be conducted in large-scale seeds classification and quality detection in modern seed industry.

Characterization of Endogenous Nanoparticles from Roasted Chicken Breasts

Emergence of endogenous nanoparticles in thermally processed food has aroused much attention due to their unique properties and potential biological impact. The aim of this study was to investigate the presence of fluorescence nanoparticles in roasted chicken breasts, elemental composition, physicochemical properties, and their molecular interaction with human serum albumin (HSA). Transmission electron microscopy analysis revealed that the foodborne nanoparticles from roasted chicken were nearly spherical with an average particle size of 1.7 ± 0.4 nm. The elemental analysis of X-ray photoelectron spectroscopy showed the composition of nanoparticles as 47.4% C, 25.8% O, and 26.1% N. The fluorescence of HSA was quenched by the nanoparticles following a static mode, and the molecular interaction of nanoparticles with HSA was spontaneous (Δ G⁰ < 0), where hydrogen bonding and van der Waals forces played an important role during HSA-nanoparticles complex stabilization through thermodynamic analysis by isothermal titration calorimetry. The principal location of the nanoparticles binding site on HSA was primarily in site I as determined by site-specific marker competition. The conformational of HSA was also changed and α-helical structure decreased in the presence of nanoparticles. Our studies revealed that fluorescent nanoparticles were produced after roasting of chicken breast at 230 °C for 30 min for the first time. The obtained nanoparticles can interact with HSA in a spontaneous manner, thus providing valuable insight into foodborne NPs as well as their effects to human albumin protein.

Dipole-dipole interactions in solution mixtures probed by two-dimensional synchronous spectroscopy based on orthogonal sample design scheme

Two-dimensional (2D) synchronous spectroscopy together with a new approach called "Orthogonal Sample Design Scheme" was used to study the dipole-dipole interactions in two representative ternary chemical systems (N,N-dimethyllformamide (DMF)/CH3COOC2H5/CCl4 and C60/CH3COOC2H5/CCl4). For the first system, dipole-dipole interactions among carbonyl groups from DMF and CH3COOC2H5 are characterized by using the cross peak in 2D Fourier Transform Infrared Radiation (FT-IR) spectroscopy. For the second system, intermolecular interaction among π-π transition from C60 and vibration transition from the carbonyl band of ethyl acetate is probed by using 2D spectra. The experimental results demonstrate that "Orthogonal Sample Design Scheme" can effectively remove interfering part that is not relevant to intermolecular interaction. Additional procedures are carried out to preclude the possibilities of producing interfering cross peaks by other reasons, such as experimental errors. Dipole-dipole interactions that manifest in the form of deviation from the Beer-Lambert law generate distinct cross peaks visualized in the resultant 2D synchronous spectra of the two chemical systems. This work demonstrates that 2D synchronous spectra coupled with orthogonal sample design scheme provide us an applicable experimental approach to probing and characterizing dipole-dipole interactions in complex molecular systems.

A Review of Optical Nondestructive Visual and Near-Infrared Methods for Food Quality and Safety

This paper is a review of optical methods for online nondestructive food quality monitoring. The key spectral areas are the visual and near-infrared wavelengths. We have collected the information of over 260 papers published mainly during the last 20 years. Many of them use an analysis method called chemometrics which is shortly described in the paper. The main goal of this paper is to provide a general view of work done according to different FAO food classes. Hopefully using optical VIS/NIR spectroscopy gives an idea of how to better meet market and consumer needs for high-quality food stuff.

Prediction of color, texture, and sensory characteristics of beef steaks by visible and near infrared reflectance spectroscopy. A feasibility study

Color, instrumental texture, and sensory attributes of steaks from 24 beef carcasses at 2, 4, 8, 14, and 21 days post mortem were predicted by visible/near infrared (visible/NIR) reflectance spectroscopy in 400-1080 nm region. Predicting the Hunter a, b, and E* yielded the coefficient of determination (R(2)) in calibration to be 0.78-0.90, and R(2) was between 0.49 and 0.55 for tenderness, Hunter L, sensory chewiness and juiciness. The prediction R(2) for tenderness was in the range of 0.22-0.72 when the samples were segregated according to the aging days. Based on partial least square (PLS) model predicted tenderness, beef samples were classified into tender and tough classes with a correct classification of 83%. Soft independent modeling of class analogy of principal component analysis (SIMCA/PCA) model of measured tenderness showed great promise in the classification of tender and tough meats with over 96% success.

Variable Selection and Updating In Model-Based Discriminant Analysis for High Dimensional Data with Food Authenticity Applications

Food authenticity studies are concerned with determining if food samples have been correctly labelled or not. Discriminant analysis methods are an integral part of the methodology for food authentication. Motivated by food authenticity applications, a model-based discriminant analysis method that includes variable selection is presented. The discriminant analysis model is fitted in a semi-supervised manner using both labeled and unlabeled data. The method is shown to give excellent classification performance on several high-dimensional multiclass food authenticity datasets with more variables than observations. The variables selected by the proposed method provide information about which variables are meaningful for classification purposes. A headlong search strategy for variable selection is shown to be efficient in terms of computation and achieves excellent classification performance. In applications to several food authenticity datasets, our proposed method outperformed default implementations of Random Forests, AdaBoost, transductive SVMs and Bayesian Multinomial Regression by substantial margins.

Near infrared and fluorescence spectroscopic methods and electronic nose technology for monitoring foods

There is a clear need for application of proper methods for measuring food quality and safety in the globalized food-webs. Numerous instrumental methods have been established in the course of the 20th century and are developing further, together with data analysis techniques, for such purposes. Among them, near-infrared and fluorescence spectroscopic methods and chemical sensor arrays called electronic noses show particular promise for rapid, non-destructive, non-invasive and cost-effective ways for assessing changes and enhancing control during processing and storage of foods. Their key advantages as analytical tools are 1) their relatively high speed of analysis, 2) the lack of a need to carry out complex sample preparation or processing, 3) their relatively low cost, and 4) their suitability for on-line monitoring or quality control. The present survey attempts to demonstrate examples from the above areas, limiting itself mainly to monitoring some quality indices which contribute to the functionality or acceptability of foods as affected by alternative processing technologies, or loss of freshness/microbial safety, or developing spoilage during storage and marketing. These instrumental methods are correlative techniques: they must be calibrated first against (traditional) reference properties, and the instrumental data are evaluated with the help of chemometric methods. Near-infrared (NIR) spectroscopy can be used in either the reflectance or the transmittance mode. NIR spectra transformed to mathematical derivatives allows subtle spectrum changes to be resolved. Selected examples from the extensive NIRS literature relate to assessment of the quality of frozen fish, predicting cooking loss of chicken patties, detecting complex physico-chemical changes of minced meat as a function of the intensity of high hydrostatic pressure treatment, comparing changes of NIR spectrometric “fingerprints” caused by gamma radiation or high pressure pasteurization of liquid egg white. Changes of NIR spectra reflect several parameters which suit the evaluation of loss of freshness, and onset of spoilage of various foods. NIR spectroscopy shows an application potential for rapid detection of bacterial or mould contamination. It may serve as a tool for detecting initial stages of mobilization processes during germination of cereal grains, or even for GMO screening. Spectrofluorometic measurements have shown potential, e.g. to monitor lipid oxidation and development of meat rancidity, to differentiate between raw and processed milks, and to monitor fish and egg freshness. Electronic noses containing chemical sensor arrays offer a rapid method for evaluation of head-space volatiles of food samples, important for characterizing quality and safety. Such gas sensors may be able to classify storage time, and determine spoilage, either earlier or at the same time as the human senses, or “sniffing out” bacterial pathogens or (toxigenic) fungal growth on certain foods. Electronic nose sensing is also a promising method for detecting quality changes of fruit- and vegetable products non-destructively. In relation to some examples to be presented in the paper, certain software developments as qualitative classification tools made by Hungarian scientists will be pointed out.

Study of the infrared spectral features of an epoxy curing mechanism

In this work, the isothermal curing process of diglycidyl ether of bisphenol A (DGEBA) cured with 4,4'-diaminodiphenylmethane (DDM) was monitored in situ by mid-infrared (MIR) and near-infrared (NIR) spectroscopy. With the help of generalized two-dimensional (2D) correlation analysis, the results obtained showed that, during curing, the change of amine and epoxy groups was simultaneous, taking place prior to the change of hydroxyl groups, followed by the change of CH(2)/CH groups, resulting from the ring-opening reaction of epoxy groups. In addition, 2D MIRxNIR hetero-spectral correlation analysis and second-derivative analysis were also employed, by means of which direct evidence of the curing mechanism could be obtained and obscure NIR band assignments in the overlapped CH combination region could be made.

Investigation of polypeptide conformational transitions with two-dimensional Raman optical activity correlation analysis, applying autocorrelation and moving window approaches

The study of conformational transitions in polypeptides is not only important for the understanding of folding mechanisms responsible for the self-assembly of proteins but also for the investigation of the misfolding of proteins that can result in diseases including cystic fibrosis, Alzheimer's, and Parkinson's diseases. Our recent studies developing two-dimensional Raman optical activity (ROA) correlation analysis have proven to be successful in the investigation of polypeptide conformational transitions. However, the complexity of the ROA spectra, and the 2D correlation synchronous and asynchronous plots, makes data analysis detailed and complex, requiring great care in interpretation of 2D correlation rules. By utilizing the 2D correlation approaches of autocorrelation and moving windows it has been possible to gain further information from the ROA spectral data sets in a simpler and more consistent way. The most significant spectral intensity changes have been easily identified, facilitating appropriate interpretation of synchronous plots, and transition phases have been identified in the moving window plots, directly relating spectral intensity changes to the perturbation.

Effect of meat cooking on physicochemical state and in vitro digestibility of myofibrillar proteins

The effect of meat cooking was measured on myofibrillar proteins from bovine M. Rectus abdominis. The heating treatment involved two temperatures (100 degrees C during 5, 15, 30, and 45 min and 270 degrees C during 1 min). Protein oxidation induced by cooking was evaluated by the level of carbonyl and free thiol groups. Structural modifications of proteins were assessed by the measurement of their surface hydrophobicity and by their aggregation state. With the aim of evaluating the impact of heat treatment on the digestive process, myofibrillar proteins were then exposed to proteases of the digestive tract (pepsin, trypsin, and alpha-chymotrypsin) in conditions of pH and temperature that simulate stomach and duodenal digestion. Meat cooking affected myofibrillar protein susceptibility to proteases, with increased or decreased rates, depending on the nature of the protease and the time/temperature parameters. Results showed a direct and quantitative relationship between protein carbonylation (p<0.01) and aggregation (p<0.05) induced by cooking and proteolytic susceptibility to pepsin. However, no such correlations have been observed with trypsin and alpha-chymotrypsin.

Two-dimensional Raman and Raman optical activity correlation analysis of the alpha-helix-to-disordered transition in poly(L-glutamic acid)

Rich and complex Raman scattering and Raman optical activity (ROA) spectra have been measured monitoring the pH induced alpha-helix-to-disordered conformational transition in poly(L-glutamic acid). Two-dimensional (2D) correlation techniques have been applied to facilitate a comprehensive analysis of these two complementary spectral sets. Synchronous contour plots have identified band assignments of alpha-helical and disordered conformations, and have revealed bands characteristic of changes in the protonation state of the polypeptide. Asynchronous plots, on the other hand, have probed the relative sequential orders of intensity changes indicating a decrease in intensity of alpha-helical bands in the backbone skeletal stretch region, followed by a subsequent decrease in intensity in the extended amide III and amide I regions, underlying the appearance of disordered structure, including poly(L-proline) II (PPII) helix. The application of a 2D correlation 'moving' window has also disclosed two distinct phases during helix unfolding in the alpha-helix-to-disordered transition, occurring at approximately pH 4.9 and approximately pH 5.2, possibly a result of the difference in helical stability between the end and central regions of the alpha-helix. This paper demonstrates the potential value of combining 2D Raman, 2D ROA and moving window correlation techniques for the detailed investigation of complex and subtle changes of secondary structure during the unfolding mechanisms of polypeptides and proteins.