Predicting amylose content, gelatinisation temperature, and RVA parameters of intact white rice using a portable NIR instrument

NIR spectroscopy has been widely applied as a rapid screening tool for quality traits in cereal breeding programs, genetic evaluation, and grain quality assessment. The objective of this study was to evaluate the feasibility of using a portable NIR instrument to predict amylose content, gelatinisation temperature (GT), and Rapid Visco Analyser (RVA) pasting parameters directly from intact white rice kernels. A total of 144 rice samples representing commercial and breeding material were analysed using a handheld NIR spectrometer operating within the 950–1600 nm spectral range. Reference measurements for amylose content, GT, and RVA parameters were obtained using standard laboratory methods. PLS regression models were developed following second derivative spectral preprocessing and dataset partitioning using the Kennard–Stone algorithm. Among the evaluated traits, GT showed the strongest predictive performance, achieving coefficients of determination of R 2 cv = 0.79 and R 2 p = 0.80, with RPD values exceeding 2.0, indicating suitability for screening applications. Moderate predictive capability was obtained for peak and trough viscosity, while amylose content and other RVA parameters exhibited limited quantitative prediction under external validation. The results demonstrate that portable NIR spectroscopy can capture spectral information related to starch–water interactions governing thermal transitions in rice, despite measurements being performed on intact kernels. Although complete replacement of laboratory-based RVA analysis remains challenging, portable NIR instruments provide a rapid, non-destructive approach for preliminary screening and classification of rice quality traits within breeding and industrial workflows.