Computational Molecular Biology 2025, Vol.15, No.5, 235-244 http://bioscipublisher.com/index.php/cmb 240 Figure 2 (A) Tissue collection and global analysis of the cattle transcriptomes. We explored expression specificity patterns by multiple analyses (differentially expressed gene analysis, stage-/tissue-specific expression, co-expression network analysis, functional enrichment, deconvolution analysis, etc.) using 108 samples from three gastrointestinal tissues (rumen, duodenum and colon) from eight lactation periods of dairy cattle. (B) The Principal Component Analysis (PCA) of three tissues across eight stages. Tissues are represented by different colors, whereas the size of the dots indicates different lactation stages. (C) The number of genes differentially expressed between adjacent stages for three tissues. (D) The number of genes differentially expressed between dry and other stages for three tissues (Adopted from Gao et al., 2024) 6.3 Identification and validation of candidate genes To identify the core factors that truly influence milk production, relying solely on the omics level is far from sufficient. Genes related to milk protein synthesis, such as CSN2 and STAT5A, have been involved in past studies (Zheng et al., 2025). But more confirmation was actually achieved through the superimposed analysis of GWAS, QTL mapping and RNA sequencing. For instance, classic genes such as DGAT1, GHR, and ABCG2 have recurred in almost all studies related to milk production. Recently noticed ones such as EFNA1, ERBB3 and CIDEAhave only gradually demonstrated their potential in regulating milk components through the integration of multiple omics. Furthermore, variations such as those in the 5’utr region, splicing sites, and even mammary mammary specific enhancers are also considered to account for a considerable proportion of trait variations (Križanac et al., 2025). To verify the functions of these genes, it is generally necessary to examine whether they overlap with QTL, whether there is functional annotation support, and the changes in expression activity under different milk production backgrounds. 7 Application Prospects of Multi-Omics Information in Molecular Breeding 7.1 Value of candidate genes and biomarkers in genomic selection Although traditional methods can capture the major effect sites, when dealing with complex traits, they often struggle to take into account changes in non-coding regions or regulatory levels. The emergence of multi-omics has precisely filled this gap. Once functional information such as genes specifically expressed in the breast, mirnas, differential expression patterns and even epigenetic markers is incorporated into genomic selection models, the prediction results will be more stable (Silpa et al., 2021). Not only that, these multi-layered validated
RkJQdWJsaXNoZXIy MjQ4ODYzNA==