<b>Mammary Gland Multi-Omics Data Reveals New Genetic Insights into Milk Production Traits in Dairy Cattle</b>

Although many sequence variants have been discovered in cattle, deciphering the relationship between genome and phenome remains a significant challenge. In this study, we defined functional classes, including mammary-specific genes, lactation-associated genes, novel long non-coding RNAs, miRNAs, RNA editing sites, DNA methylation, histone modifications, eQTLs, and sQTLs, and estimated their contributions to genetic variation for milk production traits using 3 million variants in 23,566 Holstein bulls. Sequence variants in the 5'UTR, synonymous, and splicing regions captured more variance for milk production traits than those in other genomic regions. Variants within specific genes related to lactating mammary tissue explained more genetic variance than those in other mammary tissues. Genetic variance was enriched in these lactation-up-regulated DEGs and down-regulated DEGs with small changes. We proposed a novel strategy for selecting candidate miRNAs by identifying overlaps that exhibit a significant number of mRNA targets with negative correlations, as well as a relatively large genetic variance explained by these targets. Notably, bta-miR-193-5p and bta-miR-345-5p were confirmed to enhance concentrations of total protein, triglycerides, and β-casein, as well as promote the proliferation of mammary epithelial cells. Furthermore, we found that mammary enhancers explained more genetic variance for milk production traits than repressive regulatory elements, while minor changes in DNA methylation accounted for more variance than larger changes. Finally, we constructed a new SNP panel, which improved the reliabilities of genomic predictions by 0.22%. Dividing routine SNPs into two groups based on functional classes improved the reliabilities by 0.21%. Overall, incorporating prior biological knowledge of the mammary gland directly enhances our understanding of the genetic architecture underlying milk production and improves the reliability of genomic predictions for milk production traits.尽管已在牛中发现大量序列变异，但解析基因组与表型组（phenome）之间的关联仍是一项重大挑战。本研究界定了多类功能类别，包括乳腺特异性基因（mammary-specific genes）、泌乳相关基因（lactation-associated genes）、新型长链非编码RNA（long non-coding RNAs, lncRNA）、微小RNA（microRNAs, miRNAs）、RNA编辑位点（RNA editing sites）、DNA甲基化（DNA methylation）、组蛋白修饰（histone modifications）、表达数量性状位点（expression quantitative trait loci, eQTLs）以及剪接数量性状位点（splicing quantitative trait loci, sQTLs），并利用23566头荷斯坦公牛的300万个序列变异，评估了各类功能变异对泌乳性状遗传变异的贡献。研究发现，相较于其他基因组区域，5'非翻译区（5' untranslated region, 5'UTR）、同义变异区及剪接区域内的序列变异，可解释更多泌乳性状的表型方差。泌乳乳腺组织相关特定基因内的变异，相较于其他乳腺组织的变异，可解释更多的遗传方差。遗传方差在泌乳上调的差异表达基因（differentially expressed genes, DEGs）以及变化幅度较小的下调差异表达基因中显著富集。本研究提出了一种筛选候选微小RNA的新策略：通过识别同时满足两个条件的重叠miRNA——其一为其拥有大量呈负相关的mRNA靶标，其二为这些靶标可解释相对较多的遗传方差。值得注意的是，bta-miR-193-5p与bta-miR-345-5p被证实可提升总蛋白、甘油三酯及β-酪蛋白的浓度，同时促进乳腺上皮细胞增殖。此外，本研究发现相较于抑制性调控元件，乳腺增强子可解释更多泌乳性状的遗传方差；而DNA甲基化的小幅变化相较于大幅变化，可解释更多的遗传方差。最后，本研究构建了一套全新的单核苷酸多态性（Single Nucleotide Polymorphism, SNP）分型芯片，可将基因组预测的可靠性提升0.22%；依据功能类别将常规SNP划分为两组后，可进一步将可靠性提升0.21%。综上，直接整合乳腺的先验生物学知识，可加深我们对泌乳性状遗传结构的理解，并提升泌乳性状基因组预测的可靠性。