Improving knowledge on the causative polymorphisms or genes regulating the expression of milk quantitative and qualitative traits and their interconnections plays a major role in dairy goat breeding programs and genomic research. This information enables optimization of predictive and selective tools, to obtain better-performing animals to help satisfy market demands more efficiently. Goat milk casein proteins (α_S1, α_S2, β, and κ) are encoded by 4 loci (CSN1S1, CSN1S2, CSN2, and CSN3) clustered within 250 kb on chromosome 6. Among the statistical methods used to identify epistatic interactions in genome-wide qualitative association studies (GWAS), gene-based methods have recently grown in popularity due to their better statistical power and biological interpretability. However, most of these methods make strong assumptions about the magnitude of the relationships between SNP and phenotype, limiting statistical power. Thus, the aims of this study were to quantify the epistatic relationships among 48 SNP in the casein complex on the expression of milk yield and components (fat, protein, dry matter, lactose, and somatic cells) in Murciano-Granadina goats, to explain the qualitative nature of the SNP used to quantify the genotypes produced as a result. Categorical principal component analysis (CATPCA) was used to delimit and group the number of SNP studied depending on their implications in the explanation of milk yield and components variability. Afterward, nonlinear canonical correlation analysis was used to identify relationships among and within the SNP groups detected by CATPCA. Our results suggest that 79.65% of variability in the traits evaluated may be ascribed to the epistatic relationships across and within 7 SNP groups. Two partially overlapping groups of epistatically interrelated SNP were detected: one group of 21 SNP, explaining 57.56% of variability, and another group of 20 SNP, explaining 42.43% (multiple fit ≥ 0.1). Additionally, SNP18, 32, and 36 (CSN1S2, CSN1S1, and CSN2 loci, respectively) were the most significant SNP to explain intragroup epistatic variability (component loading > |0.5|). Conclusively, milk yield and quality may not only depend on the specific casein gene pool of individuals, but may also be relevantly conditioned by the relationships set across and within such genes. Hence, studying epistasis in isolation may be crucial to optimize selective practices for economically important dairy traits.

在调节牛奶定量和定性特征及其相互关系的致病多态性或基因方面，知识的提高在奶山羊育种计划和基因组研究中起着重要作用。该信息可以优化预测和选择性工具，以获得性能更好的动物，以帮助更有效地满足市场需求。羊奶蛋白质的酪蛋白（α _S1，α _S2，β，和κ）由4个位点（编码CSN1S1，CSN1S2，CSN2和CSN3）聚集在6号染色体上的250 kb范围内。在用于确定全基因组定性关联研究（GWAS）中的上位性相互作用的统计方法中，基于基因的方法由于其更好的统计能力和生物学解释性而最近变得越来越流行。但是，大多数这些方法都对SNP与表型之间的关系强度进行了强有力的假设，从而限制了统计能力。因此，本研究的目的是量化酪蛋白复合物中48个SNP之间的上位关系，以说明Murciano-Granadina山羊的牛奶产量和成分（脂肪，蛋白质，干物质，乳糖和体细胞）的表达。 SNP的定性性质，用于量化由此产生的基因型。分类主成分分析（CATPCA）用于界定和分组所研究SNP的数量，具体取决于它们对解释牛奶产量和成分变异性的影响。此后，使用非线性典范相关分析来识别由CATPCA检测到的SNP组之间及其内部的关系。我们的结果表明，所评估性状的79.65％的变异性可能归因于7个SNP组之间和之内的上位性关系。检测到两部分重叠的上位相关SNP：一组为21个SNP，解释了57.56％的变异性，另一组为20个SNP，解释了42.43％（多重拟合≥0.1）。此外，SNP18、32和36（非线性典型相关分析用于确定由CATPCA检测到的SNP组之间及其内部的关系。我们的结果表明，所评估性状的79.65％的变异性可能归因于7个SNP组之间和之内的上位性关系。检测到两部分重叠的上位相关SNP：一组为21个SNP，解释了57.56％的变异性，另一组为20个SNP，解释了42.43％（多重拟合≥0.1）。此外，SNP18、32和36（非线性典型相关分析用于确定由CATPCA检测到的SNP组之间及其内部的关系。我们的结果表明，所评估性状的79.65％的变异性可能归因于7个SNP组之间和之内的上位性关系。检测到两部分重叠的上位相关SNP：一组为21个SNP，解释了57.56％的变异性，另一组为20个SNP，解释了42.43％（多重拟合≥0.1）。此外，SNP18、32和36（检测到两部分重叠的上位相关SNP：一组为21个SNP，解释了57.56％的变异性，另一组为20个SNP，解释了42.43％（多重拟合≥0.1）。此外，SNP18、32和36（检测到两部分重叠的上位相关SNP：一组为21个SNP，解释了57.56％的变异性，另一组为20个SNP，解释了42.43％（多重拟合≥0.1）。此外，SNP18、32和36（CSN1S2，CSN1S1，和CSN2位点，分别）是最显著SNP解释组内上位性变异性（元件装载> | 0.5 |）。结论是，牛奶的产量和质量不仅取决于个体的特定酪蛋白基因库，而且还可能受到此类基因之间和内部的关系的制约。因此，单独研究上位性对于优化具有重要经济意义的乳制品性状的选择性实践可能至关重要。