The use of near-infrared spectroscopy (NIRS) in the study of seed quality components in plant breeding programs

The potential of near-infrared spectroscopy (NIRS) for determining acid detergent fiber (ADF), fatty acid composition, and also protein and oil content in the seed of oilseed Brassica (fam. Brassicaceae) and chickpea (Cicer arietinum L.) was assessed. Accessions of the species Indian mustard (Brassica juncea L. Czern. & Coss.), Ethiopian mustard (Brassica carinata A. Braun) and rapeseed (Brassica napus L.), and also chickpea (Cicer arietinum L.), were scanned by NIRS as intact (Brassica spp.) or ground (C. arietinum) seed, and their reference values regressed against different spectral transformations by modified partial least squares (PLSm) regression. The coefficients of determination in the external validation (r2) for the different components analyzed in Brassica ranged from 0.81 (oleic acid C18:2 in B. carinata) to 0.98 (oil in B. juncea), which characterize those equations as having from good to excellent quantitative information. The standard deviation (S.D.) to standard error of prediction ratio (RPD) and S.D. to range (RER) were variable for the different species and seed constituents, and showed values that were characteristic of equations suitable for screening purposes to select samples for more detailed chemical analysis. Also, quality components in chickpea seed were determined accurately. NIRS equations showed r2 values that ranged from 0.85 (C18:3) to 0.95 (crude protein and C22:1), and RPD and RER values that were indicative of equations suitable for screening.