Nondestructive prediction of isoflavones and oligosaccharides in intact soybean seed using Fourier transform near-infrared (FT-NIR) and Fourier transform infrared (FT-IR) spectroscopic techniques

Authors: AMANAH, HANIM - Chungnam National University; TUNNY, SALMA - Chungnam National University; MASITHOH, RUDIATI - Gadjah Mada University; CHOUNG, MYOUNG-GUN - Kangwon National University; KIM, KYUNG-HWAN - Korean Rural Development Administration; Kim, Moon; BAEK, INSUCK - Orise Fellow; LEE, WANG-HEE - Chungnam National University; CHO, BYOUNG-KWAN - Chungnam National University

Submitted to: Foods
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 1/14/2022
Publication Date: 1/16/2022
Citation: Amanah, H., Tunny, S.S., Masithoh, R., Choung, M., Kim, K., Kim, M.S., Baek, I., Lee, W., Cho, B. 2022. Nondestructive prediction of isoflavones and oligosaccharides in intact soybean seed using Fourier transform near-infrared (FT-NIR) and Fourier transform infrared (FT-IR) spectroscopic techniques. Foods. https://doi.org/10.3390/foods11020232.
DOI: https://doi.org/10.3390/foods11020232

Interpretive Summary: Soybeans contain bioactive compounds such as isoflavones and oligosaccharides—phytochemicals that benefit human health and are also useful measures in seed selection for plant breeding. However, conventional methods to measure these compounds are time-consuming and destroy the sample, and thus are precluded from use for rapid food quality screening. This research investigates the feasibility of methods based on Fourier transform near-infrared (FT-NIR) and Fourier transform infrared spectroscopy (FT-IR) combined to nondestructively predict the total isoflavone and oligosaccharide contents in soybeans using intact seeds. To develop prediction models, spectral data were first acquired for 6510 seeds from 310 soybean varieties, and then a conventional sample-destructive assay was used to measure the isoflavone and oligosaccharide contents of the seeds. The models were tested using a separate set of 1365 seeds selected randomly from 65 soybean varieties. The models demonstrated acceptable prediction results that suggest that both FT-NIR and FT-IR may be feasible as nondestructive methods for analyzing intact seeds for isoflavones and oligosaccharides, although FT-NIR prediction models performed slightly better than the FT-IR prediction models. With continued development, FT-NIR and FT-IR methods could be developed for use to the benefit of both crop breeders and food processors to improve the quality of animal feed and food for human consumption.

Technical Abstract: Soybean is an important crop for the global food and feed industry. Not only high in protein, soybean also contains bioactive compounds such as isoflavones and oligosaccharides—phytochemicals that have benefits for human health. However, conventional analytical methods to measure these compounds in foods are time-consuming and destroy the sample, and thus, for seed and bean crops, are limited to use in applications for chemical-based seed selection in the breeding industry while precluded from rapid screening applications for food quality. This research investigates the feasibility of nondestructive methods based on Fourier transform near-infrared (FT-NIR) and Fourier transform infrared spectroscopy (FT-IR) combined with multivariate analysis to predict total isoflavones and oligosaccharides in soybeans using intact seed samples. The spectral data for model development were first acquired using 6510 seeds from 310 soybean varieties, divided into calibration and validation datasets, after which reference values were obtained using a conventional sample-destructive assay. The models were then tested using a new dataset of 1365 seeds selected randomly from 65 soybean varieties. In addition, an in-depth evaluation of the instruments' abilities to determine micro-chemical soybean components was carried out by developing models for specific types of isoflavones and oligosaccharides. The best prediction results of FT-NIR for intact seed evaluation were satisfactory, showing acceptable prediction of total isoflavones and oligosaccharides (Rp2: 0.80 and 0.72, respectively) with low error (SEP: 0.33 mg/g and 0.80%, respectively). The FT-NIR prediction models also demonstrated the feasibility of predicting individual types of evaluated components, with acceptable performance values (Rp2) over 0.70. Although FT_IR prediction models for total isoflavones and oligosaccharides had slightly worse performance (Rp2: 0.73 and 0.70), they still showed some feasibility for being applied to predict the examined components. Testing results confirmed model performance by obtaining R2 and error values similar to those of the calibration model. Overall, this study successfully demonstrated the feasibility of using intact seeds to predict isoflavones and oligosaccharides in soybean using FT-NIR and FT-IR spectroscopic techniques.