Gathering Reliable Data on Malting Quality for Genetic Analysis from Barley Using near Infrared Spectroscopy

Abstract

This study applies near infrared (NIR) data for phenotyping genetic populations to determine genomic regions associated with malting quality in barley. To date, most such phenotyping has used reference testing, but this is expensive when a large number of samples must be tested. NIR is a cost-effective tool, but the precision in identifying genomic regions associated with a quality trait is dependent on the accurate collection of data. To ensure that NIR is accurate for use in genomic studies, samples from genetic populations need to be validated independently. Useful interpretive statistics for both calibration and validation data, in combination, include the standard error of prediction (SEP), the RPD value (the standard deviation of the reference data divided by the SEP) and the coefficient of determination (R² and r²). The most useful of these three statistics used in this study were RPD and SEP. We demonstrated that when using NIR-predicted protein and malt extract on wholegrain barley for quantitative trait loci (QTL) analyses, an RPD value greater than 4.00 was required to ensure significant QTL were identified. We have demonstrated that NIR is an appropriate tool to phenotype barley grain for protein content and malt extract in barley mapping studies. However, when the error in the NIR data increased, and RPD values less than 4.00 were observed, which also increased the SEP values, the number of significant QTLs decreased, and the number of spurious QTLs increased.