Near-Infrared Reflectance Spectroscopy as Tool for Chemometric Analysis in Sweet Potato Vines and Cassava Foliage

Abstract

The study was conducted to evaluate near-infrared reflectance spectroscopy as a non-invasive and environment-friendlytechnique for predicting nutrients content and digestibility of sweet potato vines  and cassava foliage. Dried vines from 13 sweet potato with 65 NIRS spectra data captures and foliage from 16 cassava varieties with 80 NIRS spectra data captures were used for the study. In the SPV, crude protein, neutral detergent fibre, acid detergent fibre, non-fibre carbohydrates, beta-carotene (250 ± 22.2 µg/100 g) and vitamin A (416 ± 37.1 IU/100 g) contents were well-predicted (p< 0.05) using NIRS with coefficients of determination (R2) ranging from 0.52 to 0.81 and root mean squared errors of prediction varying from 1.42 to 89.05 in SPV. There were no significant correlations (p> 0.05) between the beta-carotene content of the SPV and the recorded nutrient contentslike DM, CP, NDF, ADF, non-fibre carbohydrates and total digestible nutrient  values. Similarly, in the CSF, mean (± s.e.) recorded in-vitro 48 hr digestibility of DM was 57.9 ± 1.47% and NDF was 80.4 ± 1.50%. Nutrient contents and in-vitro 48 h digestibility fitted well with R2 varying from 0.57 for NDF to 0.94 for ADF while, RMSEP was about 0.35 for DM, NDF and IVDMD predictions. The IVDMD was predictable from ADF and TDN contents while, IVNDFd was predictable from NDF, ADF and TDN contents. The results suggest that, NIRS technique can be developed as a rapid and accurate tool to predict the chemical composition, nutrient digestibility of samples of sweet potato vines and cassava foliage cultivars.