ENHANCING THE UNDERSTANDING OF THE INTERRELATIONSHIPS BETWEEN MORPHOLOGICAL AND PRODUCTIVE TRAITS IN SOYBEAN USING STRUCTURAL EQUATION MODELS

Understanding the relationship between multiple traits is fundamental in soybean breeding programs because their primary goal is to maximize multiple traits simultaneously, either directly or indirectly. Typically, multi-trait models are used to get genetic parameters. However, those parameters accounts only about to the linear association between traits. In other words, this approach does not account how the interrelationships between traits are defined. Therefore, we applied structural equation modeling (SEM) to explore the interrelationship between traits related to morphology (pod thickness - PT) and productive traits (number of pods - NP, number of grains - NG, and hundred grains weight - HGW). We used a dataset containing 96 soybean individuals genotyped with 4,070 single nucleotide polymorphism (SNP) markers. The phenotypic network was modeled using the hill-climbing algorithm, and the structural coefficients were estimated using the SEM approach. According to the sign of the structural coefficient, we identified positive or negative phenotypic interrelationships. We found negative interrelationships between NG and HGW, positive interrelationships between NP and NG, and between HGW and PT.