Genetic analysis without replications: model evaluation and application in spring wheat

Genetic data collected from various plant breeding and genetic studies may not be replicated in field designs although field variation is always present. In this study, we addressed this problem using spring wheat (Triticum aestivum L.) trial data collected from two locations. There were no intralocation replications and an extended additive-dominance (AD) model was used to account for field variation. We numerically evaluated the data from simulations and estimated the variance components. For demonstration purposes we also analyzed three agronomic traits from the actual spring wheat data set. Results showed that these data could be effectively analyzed using an extended AD model, which was more comparable to a conventional AD model. Actual data analysis revealed that grain yield was significantly influenced by systematic field variation. Additive effects were significant for all traits and dominance effects were significant for plant height and time-to-flowering. Genetic effects were predicted and used to demonstrate that most spring wheat lines developed by the South Dakota State University breeding program (SD lines) exhibited good general combining ability effects for yield improvement. Thus, this study provides a general framework to appropriately analyze data in situations where field crop data are collected from non-replicated designs.