UC Davis

ABSTRACT

Leveraging genetic engineering to develop herbicide and pest tolerant crops has been tremendously successful and resulted in significant benefits to farmers since the first commercialized genes became available in the mid-1990s. This technology led to the belief that step level changes were possible for highly quantitative traits such as yield and significant investment has followed. However, commercial success of large-effect yield genes continues to remain elusive. Early generation testing in transgenic programs has focused on event level and environmental variation. The transgene’s effect across different genetic background has been largely ignored until late stage evaluation. Understanding and quantifying the importance of transgene by germplasm interactions and how to best evaluate these interactions early in a transgene development pipeline may help with identifying important yield genes both in transgenic and gene editing discovery programs.To quantify the transgene by germplasm interaction in maize, we developed ACS6 RNAi transgenic and non-transgenic paired lines in each heterotic group and crossed them to each other in a factorial crossing scheme that enabled comparisons of the transgene effect on average across all germplasm, within individual line pairs, and among transgenic lines when crossed to several non-transgenic lines, and among non-transgenic lines when crossed to several transgenic lines. These hybrids were grown across a broad set of environments. For yield, there was an overall transgene effect across all hybrid sets and differences both among transgenic lines across non-transgenic line and among non-transgenic lines across transgenic lines. Genetic backgrounds had a significant impact on the estimated value of the transgene ranging from a yield reduction of .5 Mg ha-1 to an increases of 1.1 Mg ha-1. The impact of introducing a transgene that overexpressed ARGOS8 into a breeding program was also investigated to understand transgene by family interactions, the impact of event level variation, and the impact of the transgene on selection. Overall, this transgene did not have a yield effect across locations and families. However, there was a strong transgene by germplasm by location interaction. Event level variation or interactions thereof were not detected for yield. When locations were clustered by environment type, strong transgene by germplasm interactions were detected in three of the four clusters for yield. Selection difference were impacted both by family and by environment class with enrichment of transgenic lines at the top yield end of the distribution within some families in certain environments, but the opposite case in other environment by family combinations. Even though transgene by germplasm interactions are important, evaluating transgenes across both environments and the germplasm early in the transgene evaluation pipeline would be cost prohibitive using the current isogenic testing approaches. Development of new methods to enable early evaluation is critical to implementation of early transgene by germplasm evaluation. In order to optimize the testing both from the perspective of yield trial plots required for testing and the demand on creating lines and hybrid seed for testing, three methods were investigated as an alternative to generating a significant number of isogenic lines. Doubled haploids, F2:3 lines, and bulk F3 populations were compared to determine if the methods resulted in similar estimations of transgene effects for an overexpressed Zmm28 transcription factor across families and locations. Overall, even with significant transgene by family interactions, all methods had similar estimates of the transgene effect. Thus, we recommended using the F3 bulk method both for the simplicity of line creation, which could be accomplished within a year, along with the reduced number of yield trial plots required for testing, because it only requires a single set of paired plots per F3 family per location. The complexity of transgene evaluation for quantitative traits has resulted in few commercially successful genes. We hypothesized that high transgene by germplasm interactions and lack of early testing for this interaction may be one reason for the lack of success. We found consistent transgene by germplasm interaction across all three transgenes evaluated for yield, and suggest a new, cost-effective, approach to evaluate transgene by germplasm interactions leveraging bulk F3 lines.