Device for Automating Sorting of Haploid Maize Kernels

Technology #16139

Questions about this technology? Ask a Technology Manager

Download Printable PDF

Categories
Researchers
Andrew Mark Settles
Jeffrey Lynn Gustin II
Managed By
John Byatt
Assistant Director 352-392-8929

Uses Spectroscopy to Rapidly Distinguish and Sort Haploid Seeds for Maize Breeding

Single kernel near-infrared reflectance spectroscopy (NIRS) has been refined to sort haploid kernels created for use in maize breeding. Using the methods developed by UF’s researchers should reduce labor costs associated with manual sorting of seeds and multiply 10-fold the number of seeds that could be processed per hour. The maize breeding industry uses doubled haploids to produce optimized genetic strains more rapidly. Typically, less than 10 percent of kernels have haploid embryos. These kernels must be sorted from the haploid induction crosses quickly and planted in a breeding nursery as soon as possible. Conventionally, haploid seeds are manually sorted based on the presence of a particular color marker; this process is tedious and produces just 1,000 seeds per hour if the color markers are clearly expressed; they often are not. Researchers at the University of Florida have modified a single kernel NIRS sorter to accurately identify and sort haploid kernels. This device is less expensive and more autonomous than available methods.

Application

Near-Infrared Reflectance spectroscopy (NIRS) device for the accurate sorting of haploid maize kernels

Advantages

  • Throughput speed of near-infrared spectroscopy device is 10-fold higher than that of manual sorting, reducing labor costs and minimizing the time required for sorting
  • Visible color markers are not required for haploid sorting, enabling the development and testing of new inducer lines with greater and more consistent haploid generation rates

Technology

This single-kernel near-infrared spectroscopy model is able to distinguish haploid kernels from diploid kernels in haploid induction crosses using a general statistical regression model. The near-infrared spectrometer scans each seed individually to identify composition and size changes. Haploid kernels display changes in embryo size and weight; relative oil, protein and starch content; and density. These combined changes provide a strong basis for discriminating haploid kernels from diploid kernels. Using a linear discriminant analysis (LDA) model and near infrared spectroscopy, University of Florida researchers increase sorting speed 10-fold with accurate results.