NIRS determination of non-structural carbohydrates, water soluble carbohydrates and other nutritive quality traits in whole plant maize with wide range variability

Laburpena

The aim of this work was to study the potential of near-infrared reflectance spectroscopy (NIRS) to predict nonstructural carbohydrates (NSC), water soluble carbohydrates (WSC), in vitro organic dry matter digestibility (IVOMD), organic matter (OM), crude protein (CP), neutral detergent fiber (NDF), acid detergent fiber (ADF) and starch in samples of whole plant maize with a wide range of variability. The samples were analyzed in reflectance mode by a spectrophotometer FOSS NIRSystems 6500. Four hundred and fifty samples of wide spectrum from different origin were selected out of 3,000 scanned for the calibration set, whereas 87 independent random samples were used in the external validation. The goodness of the calibration models was evaluated using the following statistics: coefficient of determination (R2), standard error of cross-validation (SECV), standard error of prediction for external validation (SEP) and the RPDCV and RPDP indexes [ratios of standard deviation (SD) of reference analysis data to SECV and SEP, respectively]. The smaller the SECV and SEP and the greater the RPDCV and RPDP, the predictions are better. Trait measurement units were g/100 g of dry matter (DM), except for IVOMD (g/100 g OM). The SECV and RPDCV statistics of the calibration set were 1.34 and 3.2 for WSC, 2.57 and 3 for NSC and 2.3 and 2.2 for IVOMD, respectively. The SEP and RPDP statistics for external validation were 0.74 and 4.7 for WSC, 2.14 and 2.5 for NSC and 1.68 and 1.6 for IVOMD respectively. It can be concluded that the NIRS technique can be used to predict WSC and NSC with good accuracy, whereas prediction of IVOMD showed a lesser accuracy. NIRS predictions of OM, CP, NDF, ADF and starch also showed good accuracy.