Comparison of Phenotypic and Genomic Selection Methods for Soybean Fatty Acids.

Making the best selections from single progeny rows into replicated field trials can often be quite challenging for traits exhibiting quantitative inheritance. Phenotypic selection (PS) strategies using measured values or visual scores can be difficult for predicting broad performance due to the complexity of genotypic and environmental influences on quantitative traits.  Genomic selection (GS) offers promise for addressing these challenges in quantitative traits, such as soybean fatty acids.  The purpose of this research was to compare GS with PS for fatty acid improvement in a soybean population consisting of 282 F5 derived recombinant inbred lines (RIL).  This population was genotyped with the SoySNP50K BeadChip, of which over 17,000 SNP were polymorphic.  To simulate progeny rows, each RIL was grown in a single plot in 2010 in Knoxville, Tennessee.  The fatty acid composition of each plot was measured using gas chromatography (GC). The RIL grown in 2010 were rated for each fatty acid using the measured GC value for PS and the estimated G-BLUP value for GS.  Each RIL was then grown in 2013 in a replicated, multi-environment field trial from which LSMEANs were obtained using measured GC data.  These LSMEANs were rank correlated to the 2010 PS and GS estimates.  For each fatty acid, the GS correlation (r=0.73-0.85, P < 0.05) significantly exceeded (P<0.05) the corresponding PS correlation (r=0.66-0.78, P < 0.05).  Additionally, high and low tail selections at 15% intensity (42 RIL) using both 2010 PS and GS estimates were evaluated for performance based on 2013 data.  Using contrast statements for comparison, RIL chosen with GS outperformed (P<0.05) those chosen by PS in 6 of the 10 comparisons, and did not differ (P>0.05) in the remaining 4 comparisons. While further studies are needed, these findings demonstrate the high potential of GS for improving soybean fatty acids.