Identification of wheat lines possessing the 1AL.1RS or 1BL.1RS wheat-rye translocation by near-infrared reflectance spectroscopy

Citation
Sr. Delwiche et al., Identification of wheat lines possessing the 1AL.1RS or 1BL.1RS wheat-rye translocation by near-infrared reflectance spectroscopy, CEREAL CHEM, 76(2), 1999, pp. 255-260
Citations number
21
Categorie Soggetti
Agricultural Chemistry
Journal title
CEREAL CHEMISTRY
ISSN journal
00090352 → ACNP
Volume
76
Issue
2
Year of publication
1999
Pages
255 - 260
Database
ISI
SICI code
0009-0352(199903/04)76:2<255:IOWLPT>2.0.ZU;2-Z
Abstract
Wheat-rye chromosomal translocations, particularly those involving the shor t arm of rye chromosome 1R, have been used during the past 25 years to inst ill resistance to plant pathogens and insects and improve the hardiness, ad aptation, and yield of wheat. Unfortunately, the presence of the 1AL.1RS or 1BL.1RS rye translocations in wheat has been shown to impart inferior doug h handling and baking characteristics. Although numerous analytical techniq ues (e.g., HPLC, monoclonal antibody tests, highperformance capillary elect rophoresis) have been developed for detecting these translocations, the com plexity of the analytical procedures restricts their use to research and an alytical laboratories. The purpose of this study was to examine the potenti al of diffuse reflectance near-infrared spectroscopy, a well-accepted techn ique in the grain industry, for detecting 1RS-containing genotypes. This re search used three independent groups of wheat samples, ranging in genetic d iversity from sister lines derived from 1RS breeding populations to commerc ial cultivars. Based on the diffuse reflectance spectra (1,100-2,500 nm) of flour, partial least squares (PLS) models, through cross-validation, exhib ited misclassification rates as low as 0%, particularly for commercial cult ivars. Misclassification rates for corresponding, but separate, test sets w ere as low as 1%. When the same modeling procedure was applied to samples o f more closely related genetic backgrounds, cross-validation misclassificat ion rates rose to 15-20%. Most problematic were samples that were heterogen eous for 1RS such as the cultivar Rawhide. incorporating heterogeneous samp les into a calibration equation improved the classification accuracy of the se samples but diminished the prediction accuracy of nonheterogeneous sampl es.