Unraveling the environmental drivers of wheat grain quality: A chemometric framework for processing NIR spectral variance

Abstract

Wheat grain quality is highly susceptible to environmental fluctuations. This study proposes a complementary two-stage chemometric framework combining Analysis of Variance-Simultaneous Component Analysis (ASCA) and ANOVA-Common Dimensions (AComDim) with near-infrared (NIR) spectroscopy to systematically evaluate spatiotemporal impacts on wheat. A total of 179 samples collected across four Austrian locations over a three-year period were analyzed. In the first stage, ASCA was utilized as a computationally efficient pre-screening tool to systematically evaluate 27 preprocessing protocols, aimed at maximizing target-factor variance and minimizing background noise. In the second stage, AComDim was implemented as the core engine for orthogonal variance decomposition. By embedding ANOVA into a multiblock framework, AComDim successfully preserved joint multivariate variations among variance blocks, overcoming the independent-block limitations of traditional ASCA. By mathematically decoupling distinct sources of variance directly from the untargeted spectral fingerprints, the results revealed that harvest year was the predominant driver of spectral variation (captured in Common Component 1, explaining 45.1% of the total variance). Furthermore, the year-by-location interaction (CC2, 14.5%) and the main effect of growing location (CC3, 14.1%) were successfully and orthogonally extracted. Although their global F -values did not meet the strict 95% statistical significance threshold, they exhibited structured, deterministic spectral patterns clearly distinguishable from random background noise. Loading analysis of broad spectral regions (e.g., 1100–1200 nm and 1350–1450 nm) highlighted environmentally induced macroscopic shifts in carbohydrates, lipids, proteins, and moisture status. This research provides a robust, high-throughput phenotyping framework for quantifying the spatiotemporal sensitivity of cereals, offering essential insights for stabilizing grain quality under changing environmental conditions.

Plant phenotyping relevance

NIRスペクトルから小麦粒の品質・環境応答を抽出する chemometric 手法が研究の中心であり、単なる品質測定ではなく、高スループット表現型解析フレームワークとして開発・適用されている。

Code and data availability

The supplied blocks describe NIR spectral acquisition and ASCA/AComDim analysis of 179 wheat samples, but contain no data availability statement, no public dataset deposit, and no author code/workflow URL. The only author URL is a ResearchGate profile, which is not a paper-specific asset. Supplementary materials (Table

No evidence-backed public reproduction asset is currently recorded.