Evaluation of Seed Composition and Agronomic Traits in Bi-Parental Glycine max By Glycine Soja Populations for Developing Genomic Prediction Models.

Abstract: Soybean [Glycine max (L.) Merr.] meal is used largely for livestock feed rations in developed countries, therefore soybean varieties with relatively high protein concentration are desirable to improve the crop value. However, over the last several decades there has been a reduction of protein concentration an increase in oil content in soybean varieties in the mid-west U.S., mainly due to the selection of higher seed yield potential.  The objective of this research is to a) characterize and estimate the phenotypic correlations among seed composition traits in unique bi-parental populations, b) estimate seed yield and seed composition phenotypic correlations, and c) develop and validate genomic prediction models for efficient selection for all measured traits. We evaluated 794 RIL’s derived from seven unique G. max by G. soja bi-parental populations during 2015 in Columbia, MO.  Maturity, lodging, shattering and disease incidence were collected in the field and seed protein, oil and sucrose concentrations were estimated using a FOSS Rapid Content Analyzer near-infrared spectrometer. The genotyping analysis is being conducted using the 6K SNP Infinium chip. The range in protein concentration was from 313 g kg^-1 to 457 g kg^-1, oil ranged from 105 g kg^-1 to 200 g kg^-1 and sucrose from 6 g kg^-1 to 63 g kg^-1. Pearson’s correlation coefficient revealed that protein concentration is negatively associated with oil content (-0.65, P ≤0.001) and sucrose (-0.57, P ≤0.001). Also, we found a positive correlation between oil content and sucrose (0.27, P ≤0.01). There are significant differences for protein, oil and sucrose across the seven populations. The highest mean values for protein were found in the populations Williams 82 x PI458536 and Williams 82 x PI464890B, for oil populations Williams 82 x PI522226 and Williams 82 x PI549046, and for sucrose populations Osage x PI479751, Osage x PI593983, and Williams 82 x PI522226. Currently, we are genotyping the bi-parental populations and conducting yield trials across four locations in Missouri during 2016, to develop and validate genomic prediction models to enhance genetic gain when utilizing wild x elite populations in a breeding program.

Keywords: protein concentration, sucrose, NIR, Glycine soja.