Animal Molecular Breeding 2024, Vol.14, No.1, 95-105 http://animalscipublisher.com/index.php/amb 96 The potential impact of genomic prediction on the livestock industry is substantial. It offers the possibility of increasing the accuracy of selection, reducing the generation interval, and enabling the selection for traits that are difficult to measure or can only be measured late in life or post-mortem (Bolormaa et al., 2013; Lopes et al., 2020). Moreover, the integration of genomic information with traditional breeding values can lead to more accurate genetic evaluations, as demonstrated in Nelore cattle for carcass and meat quality traits (Lopes et al., 2020). Despite the promise of genomic prediction, challenges remain. The accuracy of genomic breeding values (GEBVs) can vary widely between traits and breeds, and the effectiveness of genomic selection is influenced by factors such as the size of the reference population and the number of markers used (Hayes and Goddard, 2010; Bolormaa et al., 2013). Additionally, the integration of biological priors, such as gene ontology (GO) terms, into genomic prediction models has been shown to improve predictive ability for certain traits, suggesting that a better understanding of the genetic architecture can enhance genomic prediction (Fang et al, 2017). In conclusion, genomic prediction technologies represent a significant advancement in the field of animal breeding, offering a means to overcome the limitations of traditional breeding methods and to accelerate genetic progress for complex traits in livestock. As research continues to refine these technologies, their integration into breeding programs is likely to become increasingly widespread, with profound implications for the livestock industry. 1 Theoretical Foundations of Genomic Selection 1.1 Principles of genomic selection Genomic selection is a revolutionary approach in livestock breeding that leverages dense single nucleotide polymorphism (SNP) genotypes to predict the genetic merit of animals. The relationship between genetic markers and phenotypes is a cornerstone of this method, as it allows for the estimation of breeding values which are essential for the selection process (Hayes et al., 2010). The accuracy of these genomic predictions is contingent upon the genetic architecture of the trait in question, such as the number of loci affecting the trait and the distribution of their effects (Hayes et al., 2010; Swami, 2010; Kemper and Goddard, 2012). For instance, traits like coat color in Holstein cattle have been found to be influenced by a few loci of large effect, as well as many loci of small effect (Hayes et al., 2010). This contrasts with other traits, such as overall type, which are affected only by loci of small effect (Hayes et al., 2010). Hayes et al. (2010) presented a study that demonstrates the relationship between F values and genomic positions of three gene loci (Figure 1). Figure A shows the distribution of the KIT gene locus, which has a clear peak in F value at approximately 72 Mb. Figure B displays the MITF gene locus, which has a higher F value at about 32 Mb. Figure C shows the PAX5 gene locus, where there is a slightly prominent F value at about 64 Mb. These data can be used to reveal the associations between specific gene variations and phenotypic traits. Genome-wide selection is another aspect of genomic selection that involves regressing phenotypes on thousands of markers concurrently. This method has been applied in both plant and animal breeding and is critical for improving the accuracy of predictions for complex traits (Campos et al., 2013). The accuracy of genomic predictions is higher for traits with a proportion of large effects, provided that the method of analysis capitalizes on the distribution of loci effects (Hayes et al., 2010). 1.2 Statistical models and algorithms The Best Linear Unbiased Prediction (BLUP) and Bayesian models are statistical tools used to enhance genomic predictions. BLUP|GA, a variant of BLUP, has been proposed to incorporate genome-wide association study (GWAS) results into genomic predictions, thereby improving the accuracy of predictions for certain traits (Zhang et al., 2014). Bayesian models, on the other hand, assume that marker effects are random variables drawn from a specified prior distribution, which has been shown to achieve high accuracy in predicting genetic values (Kemper and Goddard, 2012).
RkJQdWJsaXNoZXIy MjQ4ODY0NQ==