The National Mungbean Improvement Program (NMIP) has been at the forefront of new statistical analyses of plant breeding data since 2003. These highly efficient tools were central to the identification of breakthrough varieties Crystal and Satin II (2008), to the development and selection of new varieties for release in 2013 and 2014, and to the transition from empirical to objective mungbean breeding.
Statistical analysis of traits such as yield and disease scores which are measured for a set of genotypes in trials conducted across multiple locations and years, is called Multi-Environment Trial (MET) data analysis. The models for MET analysis have evolved and improved over the past decade, and mungbean breeders have used the results from fitting these models to design crosses and evaluate genotypes with greater confidence.
Most recent improvements to the MET analysis incorporate pedigree information of genotypes into the linear mixed models, which allows for estimation of additive genetic effects for each genotype and their ancestors. These additive genetic effects (also called estimated breeding values, EBVs, in animal breeding) are useful for identifying the best parents to use in future crosses. Selecting parents based on their EBVs is used to produce crosses specifically directed to combining populations for high yield and disease resistance.
The total genetic effects, which are the sum of the additive and the non-additive genetic effects (estimated from the same analysis for all genotypes in the trials), can help identify superior varieties for commercial release, because it assists in the identification of the best individuals for the growing conditions.
Analysis of MET data using pedigrees has been the routine analysis used by the NMIP for the last two years. A natural consequence will be better crossing and selection decisions resulting in increased efficiency and effectiveness of the program into the future.