Malt quality profile of barley predicted by near-infrared spectroscopy using partial least square, Bayesian regression, and artificial neural network models

Authors: Ajayi, Oyeyemi; AKINYEMI, LANRE - Hampton University; ATANDA, SIKIRU - North Dakota State University; Walling, Jason; Mahalingam, Ramamurthy

Submitted to: Journal of Chemometrics
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 9/14/2023
Publication Date: 10/1/2023
Citation: Ajayi, O.O., Akinyemi, L., Atanda, S.A., Walling, J.G., Mahalingam, R. 2023. Malt quality profile of barley predicted by near-infrared spectroscopy using partial least square, Bayesian regression, and artificial neural network models. Journal of Chemometrics. e3519: 1-15. http://doi.org/10.1002/cem.3519.
DOI: https://doi.org/10.1002/cem.3519

Interpretive Summary: Barley (Hordeum vulgare L.) is the fourth most important cereal crop globally and is used for feed, malting industry, and food. The increasing demand for novel barley genotypes with the desired grain quality underscored the need to develop fast, accurate and cost-effective methods that breeders can use to aid in selection of high-performing lines. By applying simple and complex statistical methods on 1,930 near-infrared (NIR) spectral and malt quality data from seventeen locations, we evaluated the utility of NIR modeling techniques to accurately predict malt quality traits such as diastatic power (DP), alpha amylase (AA), malt extract (ME), wort protein (WP), soluble to total protein (S/T) ratio and free amino nitrogen (FAN). Although the highest prediction accuracy was obtained for WP (56%), influential wavelength regions identified could be exploited to enhance the future design of NIR prediction models for the quick identification and selection of high performing lines during malt quality improvement programs.

Technical Abstract: Due to the significant cost and time involved in identifying barley lines with superior malting quality, the malting industry is searching for accurate and rapid methods to expedite the selection of superior barley lines that meets breeder’s goals. The aim of this study was to compare partial least square regression (PLSR) with advanced statistical models (Bayesian and machine learning) and reliably assess their performance in predicting malt quality traits from near infra-red (NIR) spectral data using barley grains. Using spectral data as predictors and the malt quality traits as reference, PLSR outperformed Bayesian and PCA-ANN model for diastatic power (DP), alpha amylase (AA), malt extract (ME), wort protein (WP), soluble to total protein (S/T) ratio and free amino nitrogen (FAN). Although across all traits, WP had the best prediction performance for all models, it demonstrated low accuracy, with the best performing model, PLSR, having R2 (RPD) values of 0.55 (1.5). The influential wavelength regions identified based on the variable importance in projection (VIP) scores and coefficient estimates for PLSR and Bayesian models respectively, were comparatively similar for all malt quality traits. Based on these findings, PLSR analysis and wavelength selection techniques would enhance the future design and optimization of NIR prediction models in malt quality improvement programs.