Phenotypic characterization of sweet potato Ipomoea batatas (L.) Lam. genotypes in relation to prediction of chemical quality constituents by NIRS equations

This study evaluates the potential of NIRS (near-infrared reflectance spectroscopy) to predict the major constituents (starch, sugars, cellulose, proteins and minerals) of the sweet potato root. Overall, 240 accessions were morphologically described, chemically analysed and their NIR spectra recorded. No correlations were observed between aerial and underground traits, and between morphological traits and major constituents. Calibration equations, developed on 190 accessions, showed high explained variances in cross-validation (r2cv) for starch (0.82), sugars (0.91), proteins (0.89) and minerals (0.74) but no response for cellulose (0.21). The predictions were tested on an independent set of 50 randomly selected accessions. The r2pred values for starch, sugars and proteins were, respectively, of 0.71, 0.82 and 0.87 with ratios of performance to deviation (RPD) of 2.11, 2.29 and 2.93. New calibration equations developed on 240 accessions showed improved RPD values for starch (2.65), sugars (2.75), cellulose (1.58), proteins (3.47) and minerals (2.34), indicating that larger sets could improve prediction. NIRS could be used in sweet potato breeding programmes to predict starch, sugars and proteins contents in the roots.

Saved in:
Bibliographic Details
Main Authors: Lebot, Vincent, Ndiaye, André, Malapa, Roger
Format: article biblioteca
Language:eng
Subjects:F30 - Génétique et amélioration des plantes, Q04 - Composition des produits alimentaires, U30 - Méthodes de recherche, F60 - Physiologie et biochimie végétale, F50 - Anatomie et morphologie des plantes, Ipomoea batatas, igname, qualité, composition chimique, génotype, technique de prévision, spectroscopie infrarouge, teneur en glucides, teneur en éléments minéraux, teneur en protéines, amidon, cellulose, partie souterraine, partie aérienne, légume racine, phénotype, morphologie végétale, http://aims.fao.org/aos/agrovoc/c_3937, http://aims.fao.org/aos/agrovoc/c_8478, http://aims.fao.org/aos/agrovoc/c_6400, http://aims.fao.org/aos/agrovoc/c_1794, http://aims.fao.org/aos/agrovoc/c_3225, http://aims.fao.org/aos/agrovoc/c_3041, http://aims.fao.org/aos/agrovoc/c_28568, http://aims.fao.org/aos/agrovoc/c_1298, http://aims.fao.org/aos/agrovoc/c_4848, http://aims.fao.org/aos/agrovoc/c_6251, http://aims.fao.org/aos/agrovoc/c_7369, http://aims.fao.org/aos/agrovoc/c_1423, http://aims.fao.org/aos/agrovoc/c_32357, http://aims.fao.org/aos/agrovoc/c_32353, http://aims.fao.org/aos/agrovoc/c_6647, http://aims.fao.org/aos/agrovoc/c_5776, http://aims.fao.org/aos/agrovoc/c_13434, http://aims.fao.org/aos/agrovoc/c_5159,
Online Access:http://agritrop.cirad.fr/560005/
http://agritrop.cirad.fr/560005/1/document_560005.pdf
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:This study evaluates the potential of NIRS (near-infrared reflectance spectroscopy) to predict the major constituents (starch, sugars, cellulose, proteins and minerals) of the sweet potato root. Overall, 240 accessions were morphologically described, chemically analysed and their NIR spectra recorded. No correlations were observed between aerial and underground traits, and between morphological traits and major constituents. Calibration equations, developed on 190 accessions, showed high explained variances in cross-validation (r2cv) for starch (0.82), sugars (0.91), proteins (0.89) and minerals (0.74) but no response for cellulose (0.21). The predictions were tested on an independent set of 50 randomly selected accessions. The r2pred values for starch, sugars and proteins were, respectively, of 0.71, 0.82 and 0.87 with ratios of performance to deviation (RPD) of 2.11, 2.29 and 2.93. New calibration equations developed on 240 accessions showed improved RPD values for starch (2.65), sugars (2.75), cellulose (1.58), proteins (3.47) and minerals (2.34), indicating that larger sets could improve prediction. NIRS could be used in sweet potato breeding programmes to predict starch, sugars and proteins contents in the roots.