Legume Genomics and Genetics 2024, Vol.15, No.6, 270-279 http://cropscipublisher.com/index.php/lgg 275 Figure 2 The GmNFYB17 transgenic soybean lines under drought treatment (Adopted from Sun et al., 2022) Image caption: (A) Morphology of transgenic and non-transgenic plants under drought conditions. Water was withheld for 15 d, and then plants were re-watered for 7 d. G16, G18 and G26 are transgenic lines; CK is soybean DN50; (B,C)The leaf relative water content (RWC) and leaf damage of transgenic lines. G16, G18, G26 and non-transgenic control (CK) during the well-watered, drought and re-watered stage; (D-F) Comparison of physiological and biochemical indicators (MDA, SOD, Proline) between transgenic and non-transgenic plants. *: p-value ≤ 0.05; **: p-value ≤ 0.01 (Adopted from Sun et al., 2022) 7 Challenges and Future Prospects 7.1 Technical and computational challenges in integrating GWAS and GS Integrating Genome-Wide Association Studies (GWAS) and Genomic Selection (GS) in soybean breeding presents several technical and computational challenges. One of the primary issues is the need for advanced statistical methods to accurately detect Quantitative Trait Loci (QTL) and their interactions. Traditional GWAS methods often focus on main effects and may miss significant interactions between QTL, which are crucial for complex traits like yield and seed quality (Yoosefzadeh-Najafabadi et al., 2023). Additionally, the computational power required to handle large datasets and perform complex analyses is substantial. Machine learning algorithms, such as Support Vector Regression (SVR) and Random Forest (RF), have shown promise in improving the accuracy of QTL detection, but they also demand significant computational resources and expertise (Yoosefzadeh-Najafabadi et al., 2021a; Yoosefzadeh-Najafabadi et al., 2021b). 7.2 Cost and resource considerations for breeding programs The integration of GWAS and GS into breeding programs is resource-intensive. The costs associated with genotyping, phenotyping, and data analysis can be prohibitive, especially for smaller breeding programs. High-throughput phenotyping technologies and the generation of large-scale genomic data require substantial financial investment and technical infrastructure (Sandhu et al., 2022). Moreover, the need for continuous updates to computational tools and methods to keep pace with advancements in the field adds to the overall cost. Efficient allocation of resources and strategic planning are essential to maximize the return on investment in these technologies.
RkJQdWJsaXNoZXIy MjQ4ODYzNA==