Phenotypic and genotypic variation within soybean cultivars

Abstract

Soybean plays a fundamental role in the Brazilian agribusiness, mainly because the country is the world's largest producer of this crop. Consequently, there is a constant pursuit of homozygous and pure genotypes in commercial soybean breeding programs. However, genetic homogeneity and purity do not remains stable over the medium and long terms due to the notable instability and flexibility of the soybean genome, which leads to intracultivar variability. Thus, the objective of this study was to investigate intracultivar variability in soybean lines, using both phenotypic and genotypic data. Efforts were also made to compare different strategies for ranking soybean progenies and to select the most promising lines for traits exhibiting variability. During the 2017/2018 crop season, 47 plants from six soybean cultivars were used to set the genetic treatments. These progenies, along with control, were subjected to trials across two subsequent crop seasons, 2018/2019 and 2019/2020. Phenotypic traits such Grain Yield (YIELD), Full Maturity (FM), Days to Flowering (DF), and Plant Height (PH) were estimated. Additionally, genotyping was performed for 288 genotypes using a chip with 1329 SNP markers. The phenotypic data were analyzed using mixed models to evaluate variance, heritability, accuracy, and coefficient of variation, as well as to estimate a multi-trait index. Spearman's rank correlation (ρ) and the coincidence index (IC) were computed using predicted genetic values from different analytical strategies like BLUP and GBLUP. This study reveals the existence of both phenotypic and genotypic intracultivar variation among the evaluated cultivars. The degree of variation observed differs among cultivars. The potential for both selection and discarding is present across all evaluated traits individually and also when employing a multi-trait index. However, the correlation between genotypic and phenotypic predictions for the purpose of selection is low. However, this correlation is higher for discarding, suggesting genotypic data is more applicable for identifying and discarding undesirable progenies. Furthermore, traits such plant height (PH) and days to flowering (DF) showed high consistency between genotypic and phenotypic predictions, indicating they are efficient in the selection and discarding processes within a multi-trait index. These results emphasize the importance of considering trait-specific predictability when applying genomic information to breeding strategies and validate the multi-trait index as a useful tool for improving the efficiency of breeding programs.