Computational Molecular Biology 2025, Vol.15, No.5, 235-244 http://bioscipublisher.com/index.php/cmb 236 related to milk yield and milk composition, clarify the biological pathways and cell types involved in milk synthesis, and evaluate the potential of multi-omics integration in enhancing the reliability of genomic selection. By bridging the gap between genotypes and phenotypes, this study aims to significantly accelerate genetic improvement, optimize breeding strategies, and ensure the long-term sustainability of dairy cow production systems. 2 Overview of the Genetic Basis of Milk Production Traits in Dairy Cattle 2.1 Major milk production traits and their genetic parameters In terms of milk production by dairy cows, indicators such as milk yield (MY), milk fat production (FY), and milk protein production (PY) have long been studied as the main traits. Milk fat percentage and milk protein percentage are also often included in the improvement targets. Although these traits are regulated by multiple genes, their heritability is not uniform. For example, the estimated value of MY is usually between 0.12 and 0.35, and the same is true for FY and PY. However, the heritability of milk fat percentage and milk protein percentage is generally lower, usually within the range of 0.06 to 0.16 (Lu et al., 2022). Interestingly, these yield-related traits are often positively correlated with each other, but when yield and percentage traits are considered together, a negative correlation sometimes occurs. Incidentally, MY repeatability reaches 0.47, that is to say, although genes play a role, the influence of the environment is also considerable (Fotso-Kenmogne et al., 2025). 2.2 Key genes and QTL regions influencing milk yield Many people think that "high milk production" depends on the "average strength" of the entire genome. But in fact, key genes like DGAT1, which are frequently mentioned, show stable effects on milk production (MY), milk fat (FY), and milk protein (PY), especially on chromosome 14 (Kim et al., 2021; Tahir et al., 2025). However, do not overlook other genes, such as MGST1, GHR, ABCG2, CSN1S1, as well as the newly emerged candidate genes in recent years, like PDE4B, ANO2 and NCOA6. Nowadays, many QTLS have almost been labeled on all autosomes, especially the bunch on chromosome 14 that is closely related to milk fat production and fat percentage (Bekele et al., 2023; Cortes-Hernandez et al., 2025). Meanwhile, multi-omics data began to come into play. For instance, when mammary mammary specific expression, non-coding RNA, regulatory factors and epigenetic modifications were all taken into account, it was found that more key variations could be identified and the prediction accuracy improved significantly (Cai et al., 2025; Križanac et al., 2025). 2.3 Interaction between genetic background and environmental factors on milk performance Ultimately, genes are not the only variable. The performance of cattle often depends on the weather, feeding conditions and the nutrition of the feed. The interaction between genotype and environment (GxE) may lead to different manifestations of the same genotype under different conditions, and even the estimated heritability may vary as a result (Mancin et al., 2023). For instance, in hot or malnourished environments, the heritability of MY may be lower than expected, while certain genotypes may perform better under these conditions (Martins et al., 2025). In addition, the traits of reproductive capacity and calving interval themselves have low heritability and are highly sensitive to management impacts. While "survival traits" such as stress resistance and heat resistance are not necessarily directly related to milk production, they are now also included in the selection index to improve adaptability (Oloo et al., 2025). Overall, if the GxE effect is not taken into account, the selected cattle may have "good paper performance", but they will "fail" once the environment changes. 3 Types and Characteristics of Multi-Omics Data 3.1 Genomic data (SNP, QTL, GWAS) Genetic variations related to milk production were first explored through genomic data. In fact, not only SNPS in the coding region, but also in many non-coding and regulatory regions can be detected through high-throughput typing and whole-genome sequencing (Buaban et al., 2021; Ristanic et al., 2024; Venkatesan et al., 2025). It is not surprising that genes like DGAT1, GHR and CTNNA3, which are "familiar faces", have been repeatedly captured in GWAS or QTL analyses of different varieties. However, to more precisely identify these sites, a sufficient number of samples are needed and the annotations must be detailed. Especially for those variations located in regulatory regions with unknown functions, it is even more necessary to integrate functional information such as
RkJQdWJsaXNoZXIy MjQ4ODYzNA==