A critical review on quantification of food bioactive components by NIR and chemometrics: Principles, advances, and challenges

Authors: TIAN, WENFEI - Chinese Academy Of Agricultural Sciences; LI, YONGHUI - Kansas State University; GUZMAN, CARLOS - Universidad De Cordoba; IBBA, MARIA - International Maize & Wheat Improvement Center (CIMMYT); Tilley, Michael - Mike; WANG, DONGHAI - Kansas State University; HE, ZHONGHU - International Maize & Wheat Improvement Center (CIMMYT)

Submitted to: Comprehensive Reviews in Food Science and Food Safety
Publication Type: Literature Review
Publication Acceptance Date: 9/19/2023
Publication Date: 9/23/2023
Citation: Tian, W., Li, Y., Guzman, C., Ibba, M., Tilley, M., Wang, D., He, Z. 2023. A critical review on quantification of food bioactive components by NIR and chemometrics: Principles, advances, and challenges. Comprehensive Reviews in Food Science and Food Safety. https://doi.org/10.1016/j.jfca.2023.105708
DOI: https://doi.org/10.1016/j.jfca.2023.105708

Interpretive Summary: The health-promoting effects of bioactive components in food are widely recognized. The contents of bioactive components (nutraceutical value) are becoming an important parameter in evaluating overall quality and market preference of food products. The nutraceutical value of foods can be described by in vitro methods such as total phenolic content (TPC), total flavonoid content (TFC), total anthocyanins content (TAC) and antioxidant activity assays such as oxygen radical absorbance capacity (ORAC). Specific bioactive compounds such as vitamin C, ferulic acid, gallic acid, caffeine and catechin can be extracted and then quantified by liquid chromatography. These conventional wet-chemistry methods are destructive to samples, costly in terms of time and labor, require experienced laboratory skills, and generate hazardous chemical waste. Such disadvantages make conventional methods unsuitable when large numbers of samples need to be analyzed in a short timeframe. Consequently, it is important to develop rapid, low-cost, chemical-free methods for rapid quantification of bioactive components to meet future demands for nutrition and sustainability. Chemometrics is the discipline that connects near-infrared (NIR) spectral information to chemical properties of interest using statistical, mathematical, and other methods to provide maximum relevant chemical information by analyzing chemical data. Coupling of chemometrics and NIR spectroscopy has been widely used for qualitative analysis, product authentication, detection of adulteration, and discrimination of geographical origin. More importantly, NIR is an advancing technique for rapid quantification of bioactive components in food since it is non-destructive, easy to use, cost-effective and eco-friendly. This manuscript provides a comprehensive and critical review emphasizing potential pitfalls in developing NIR methods for food analysis.

Technical Abstract: With increasing consumer demand for health-promoting foods the contents of bioactive components are becoming an important aspect in evaluation of food quality. Conventional wet-chemistry methods for quantitative analysis are time-consuming, costly, and generate hazardous waste. A combination of near-infrared (NIR) spectroscopy and chemometrics is a rapidly advancing way of assessing food quality. The success of quantitative NIR models require in-depth knowledge spanning multiple disciplines. Various common issues and pitfalls that may occur during model development have not been widely recognized. This review presents the key principles of related disciplines and a general workflow for NIR model development including sample selection, data collection, outlier detection, spectral preprocessing, model calibration, and model evaluation and validation. Some common issues and potential pitfalls that can occur during model development have not been widely recognized. Findings, highlights, implications, and limitations from recent studies are critically discussed. The major challenges of NIR technology regarding sample representativeness, comparison of algorithms and data interpretation are systemically evaluated. Future perspectives on NIR method development, advantages of miniatured NIR systems and potential NIR applications are described and assessed. Greater research collaboration and data sharing would be highly beneficial in developing advanced powerful NIR methods for wider use.