Animal Molecular Breeding 2024, Vol.14, No.3, 239-251 http://animalscipublisher.com/index.php/amb 241 Bayesian methods, on the other hand, provide more flexibility by allowing different prior distributions for marker effects, which can be particularly useful when dealing with traits influenced by large-effect loci. More recently, machine learning algorithms have been explored for GS, offering the potential to capture complex interactions between markers that traditional models might miss. The choice of model depends on the specific breeding objectives, the genetic architecture of the traits in question, and the available data. Advances in computational power and statistical methodologies continue to drive the development of more accurate and efficient genomic prediction models, further enhancing the potential of GS in livestock breeding (Ibtisham et al., 2017). 2.4 Marker-assisted selection vs. genomic selection Marker-assisted selection (MAS) and genomic selection (GS) are both used to enhance the genetic improvement of livestock, but they differ significantly in their approach and effectiveness. MAS relies on a limited number of markers linked to specific traits, making it suitable for traits controlled by a few major genes (Meuwissen et al., 2016). However, its effectiveness diminishes for complex traits governed by many genes with small effects. In contrast, GS uses high-density SNP data covering the entire genome, allowing for the simultaneous consideration of all genetic markers. This comprehensive approach enables GS to capture the total genetic variance, including contributions from small-effect loci that MAS might miss. As a result, GS provides more accurate predictions of breeding values and can be applied to a broader range of traits. The ability of GS to reduce the generation interval and increase the rate of genetic gain makes it a superior tool for modern livestock breeding programs. The transition from MAS to GS marks a significant evolution in genetic selection strategies, reflecting the advancements in genomic technologies and the growing understanding of complex trait genetics (Priyadarshini et al., 2017). 3 Advances in Genomic Selection The field of genomic selection (GS) in livestock breeding has evolved significantly over the past decade, driven by technological innovations, the integration of big data and machine learning, improvements in genomic prediction accuracy, and enhanced breeding program efficiency. These advances have not only improved the accuracy of selection but also accelerated the rate of genetic gain, making GS an essential tool in modern livestock breeding. 3.1 Technological innovations in genomic sequencing Technological advancements in genomic sequencing have played a critical role in the evolution of GS. The development of high-throughput sequencing technologies has enabled the generation of vast amounts of genomic data at a reduced cost, making it feasible to implement GS on a large scale. The introduction of next-generation sequencing (NGS) technologies, such as whole-genome sequencing and genotyping-by-sequencing, has revolutionized the ability to capture genetic variation across the entire genome. These innovations have led to the identification of single nucleotide polymorphisms (SNPs) that serve as the basis for GS, allowing for more accurate predictions of breeding values. Additionally, advances in sequencing technologies have improved the resolution of genomic data, enabling the detection of rare variants and structural variations that contribute to complex traits in livestock. The combination of these technological improvements has resulted in a more comprehensive understanding of the genetic architecture of economically important traits, thereby enhancing the effectiveness of GS in breeding programs (Meuwissen et al., 2016). 3.2 Integration of big data and machine learning The integration of big data and machine learning into GS has significantly enhanced the ability to predict complex traits in livestock. The vast amounts of genomic, phenotypic, and environmental data generated through modern breeding programs require advanced analytical tools to extract meaningful insights. Machine learning algorithms, such as random forests, support vector machines, and deep learning models, have been increasingly applied to GS to improve prediction accuracy.
RkJQdWJsaXNoZXIy MjQ4ODYzNA==