Combining partial least squares (PLS) discriminant analysis and Rapid Visco Analyser (RVA) to classify barley samples according to year of harvest and locality

Abstract

The aim of this study was to evaluate the usefulness of the Rapid Visco Analyser (RVA) instrument combined with pattern recognition methods as tools to differentiate commercial barley samples from two South Australian localities and three harvests. Principal component analysis (PCA), partial least squares discriminant analysis (PLS-DA) and stepwise discriminant analysis were applied to classify samples based on the RVA profiles using full cross validation (leave-one-out) as the validation method. The PLS-DA models correctly classify 96.3 and 97.8 % of the barley samples according to harvest and locality, using the profiles generated by the RVA instrument. Analysis and interpretation of the eigenvectors and loadings from the PCA or PLS-DA models developed verified that the RVA profiles contain relevant information related to starch pasting properties that allows sample classification. These results suggest that RVA coupled with PLS-DA holds necessary information for a successful classification of barley samples sourced from different localities and harvests.