Classification of individual waxy maize kernels using single kernel near infrared reflectance spectroscopy

Abstract

Two forms of starch are present in maize kernels, amylose and amylopectin. Waxy maize kernels are almost entirely composed of amylopectin, with little to no amylose present. The waxy1 locus in maize encodes an enzyme that synthesizes amylose, and mutations inactivate this enzyme. The lack of amylose in waxy maize has led to this trait having distinct properties useful for many industries. Breeders may be highly interested in incorporating this trait in their programs; however, the identification of waxy maize kernels is time and labor intensive. Near infrared reflectance spectroscopy is a fast and efficient method that could offer a rapid throughput identification of waxy maize kernels. In this project, maize samples were collected, scanned in a single kernel NIR instrument, and chemically analyzed for the waxy trait. Linear discriminant analysis was conducted to develop a classification equation capable of predicting waxy maize kernels. Results showed a greater than 90% accuracy at correctly classifying kernels as their true states. A prediction equation was successfully prepared for use to predict whether a given kernel is waxy based on its individual NIR spectrum. This project will benefit breeders looking to incorporate the waxy trait in their maize breeding programs using a rapid, non-destructive identification method.