Assessing hyperspectral and metabolomic methodologies for herbicide resistance detection in Montanan Avena fatua

Abstract

Producers spend millions of dollars each year combating herbicide resistant wild oat, often requiring high rates of application to successfully control populations. We used recombinant inbred lines to assess the viability of both untargeted metabolomics and hyperspectral imagery as potential diagnostic tools to identify resistant biotypes of wild oat. Untargeted metabolomics identified several features consistently more abundant in susceptible than resistant plants between both field-collected biotypes and their derived recombinant lines. Identification of these metabolites could provide insights into mechanisms and lead to development of resistance screening assays. Hyperspectral imagery was able to discriminate between field-collected resistant and susceptible strains, though it struggled to discriminate among similar resistance phenotypes. Within the recombinant lines, hyperspectral approaches could not predict resistance, suggesting that different computational methods are necessary to predict resistance. These insights into pre-spraying resistance detection methods are invaluable for directing future research in herbicide resistance detection in wild oat and other resistant weeds.