Oil Content is Superior to Oil Mass for Identification of Haploid Seeds in Maize Produced with High-Oil Inducers

Rapid production of inbred lines by in vivo haploid induction has become a routine method in maize breeding. Limitations of the present method can be overcome by use of inducers with high seed-oil content for discrimination of haploid (H) from diploid crossing (C) seeds. With regard to the optimum application of this new approach, our goal was to find the most accurate discrimination method comparing the directly measured parameters oil mass (OM) and seed weight (SW) with the derived parameter oil content (OC). For this, we compared the distribution properties of H and C seeds from 11 induction crosses with high-oil inducers for all three traits. The distributions of H and C seeds overlapped almost completely for SW but overlapped less for OC than for OM owing to a much smaller coefficient of variation in both fractions. Thus, the false discovery rate and false negative rate for sorting of H and C seeds were generally smaller for OC than for OM. The optimal bivariate discrimination function on the basis of OM and SW yielded similar error rates as a univariate discrimination function on the basis of OC alone. In conclusion, we recommend to measure seeds from induction crosses produced with high-oil inducers for SW in addition to OM and use OC as a simple and robust criterion for sorting of H and C seeds.