Prediction of Retinol in Fortified Maize Flour using Fourier Transform - Near Infrared Spectroscopy

• Published in: Current research in nutrition and food science 2024-04, Vol.12 (1), p.384-396
• Main Authors: Chepkoech, Brenda, N. Wafula, Elizabeth, Sila, Daniel N., N. Orina, Irene
• Format: Article
• Language: English
• Subjects: Brand names; Calibration; Compliance; Corn; Flour; Fourier transforms; High performance liquid chromatography; Infrared spectroscopy; Least squares method; Liquid chromatography; Near infrared radiation; Nutrient deficiency; Particle size; Performance measurement; Vitamin A
• ISSN: 2347-467X; 2322-0007
• DOI: 10.12944/CRNFSJ.12.1.31

Food fortification is one strategy for addressing micronutrient deficiencies among the population groups at risk. Non-compliance with fortification standards hinders the success of fortification programs. This is due to a lack of techniques to rapidly check the amounts of the added fortificants. Fourier transform - near-infrared (FT-NIR) spectroscopy is a fast and reliable technique that would be used to ensure adherence to requirements. This study aimed to investigate the potential of using FT-NIR spectroscopy to predict the amount of retinol in fortified maize flour. 150 fortified maize flour samples were used in this study. Partial least squares regression (PLS-R) was used to build calibration models based on the retinol reference values obtained by high-performance liquid chromatography (HPLC), and fortified maize flour NIR spectra acquired from the FT-NIR spectrophotometer. Two calibration models were developed to predict retinol above and below 1.0 mg/kg. The performance metrics of model one developed to predict retinol < 1.0 mg/kg were: R2c = 0.81, RMSEE = 0.08, RPD = 2.29 and R2v = 0.82, RMSEP = 0.09, RPD = 2.07 for the calibration and validation, respectively. The second model developed to predict retinol ≥ 1.0 mg/kg had the following performance metrics: R2c = 0.93, RMSEE = 0.16, RPD = 3.58 and R2v = 0.81, RMSEP = 0.22, RPD = 2.43 for the calibration and validation, respectively. Overall, the findings demonstrated that FT-NIR spectroscopy can be utilised to reliably predict retinol levels in fortified maize flour samples. FT-NIR spectroscopy, by replacing time-consuming and laborious wet chemistry laboratory procedures, has the potential to be used for rapid regulatory monitoring of fortification compliance for a large number of samples.