Balancing genomic selection efforts for allogamous plant breeding programs

Genomic selection (GS) is fundamentally a statistical genetics technique, which encourages scientists to develop robust models for this purpose. However, the application of GS is not confined to mathematical theory alone; it entails a meticulous evaluation of its practicality and applicability, particularly across generations of crossbreeding and in the strategic management of base-populations used for model calibration. While costs have diminished, it remains a substantial investment, notably due to the dollar pricing of each breeding sample. To ensure the efficiency of this technology, foresight in planning is imperative, taking into account available data, those to be acquired, and the quality of SNP and phenotypic data. Maintaining focus on the base population that will endure throughout the selection cycles of the program is paramount (given that GS models are inherently linked to relatedness among individuals). Selection strategies encompassing both additive and non-additive effects are necessary. Still, they must be applied judiciously, considering the phase of the program, be it for the development of lines, hybrids, or both. The complexity of models should be managed with prudence, considering their routine applicability; for instance, a predictive artificial intelligence model may not always be the unequivocal choice. Furthermore, it is wise to consider that in some cases, a simple pedigree-based model may deliver results as effective and more cost-efficient than GS. However, when kinship information is limited or absent, this is where genomics reveals one of its greatest advantages. Genomic models possess a unique elegance, and those who employ them are at the forefront of crop biotechnology advancement.