Over the last years, genome-wide association studies (GWAS) based on imputed whole-genome sequences (WGS) have been used to detect quantitative trait loci (QTL) and highlight candidate genes for important traits. However, in general this approach does not allow to validate the effects of candidate mutations or determine if they are truly causative for the trait(s) in question. To address these questions, we applied a two-step, within-breed GWAS approach on 15 traits (5 linked with milk production, 2 with udder health, and 8 with udder morphology) in Montbéliarde (MON), Normande (NOR), and Holstein (HOL) cattle. We detected the most-promising candidate variants (CV) using imputed WGS of 2515 MON, 2203 NOR, and 6321 HOL bulls, and validated their effects in three younger populations of 23,926 MON, 9400 NOR, and 51,977 HOL cows. Bull sequence-based GWAS detected 84 QTL: 13, 10, and 30 for milk production traits; 3, 0, and 2 for somatic cell score (SCS); and 8, 2 and 16 for udder morphology traits, in MON, NOR, and HOL respectively. Five genomic regions with effects on milk production traits were shared among the three breeds whereas six (2 for production and 4 for udder morphology and health traits) had effects in two breeds. In 80 of these QTL, 855 CV were highlighted based on the significance of their effects and functional annotation. The subsequent GWAS on MON, NOR, and HOL cows validated 8, 9, and 23 QTL for production traits; 0, 0, and 1 for SCS; and 4, 1, and 8 for udder morphology traits, respectively. In 47 of the 54 confirmed QTL, the CV identified in bulls had more significant effects than single nucleotide polymorphisms (SNPs) from the standard 50K chip. The best CV for each validated QTL was located in a gene that was functionally related to production (36 QTL) or udder (9 QTL) traits. Using this two-step GWAS approach, we identified and validated 54 QTL that included CV mostly located within functional candidate genes and explained up to 6.3% (udder traits) and 37% (production traits) of the genetic variance of economically important dairy traits. These CV are now included in the chip used to evaluate French dairy cattle and can be integrated into routine genomic evaluation.

中文翻译：

---

奶牛产奶，乳房健康和乳房形态的候选变体的确认作用

在过去的几年中，基于推算全基因组序列（WGS）的全基因组关联研究（GWAS）已用于检测定量性状基因座（QTL）并突出显示重要性状的候选基因。但是，一般而言，这种方法不允许验证候选突变的影响或确定它们是否真正引起了所讨论的性状。为了解决这些问题，我们在蒙贝利亚德（MON），诺曼德（NOR）和15个性状上采用了两步育种的GWAS方法（15个与牛奶生产相关，2个与乳房健康相关，8个与乳房形态相关）。荷斯坦（HOL）牛。我们使用估算的2515 MON，2203 NOR和6321 HOL公牛的WGS检测最有前途的候选变异（CV），并验证了它们在23,926 MON，9400 NOR和51,977 HOL牛的三个较年轻种群中的作用。基于公牛序列的GWAS检测到84个QTL：13、10和30的牛奶生产性状。体细胞评分（SCS）的3、0和2；乳房形态性状分别为8、2和16，分别在MON，NOR和HOL中。在三个品种中共有五个影响牛奶生产性状的基因组区域，而在两个品种中有六个基因区域（两个为生产性，四个为乳房形态和健康性状）。在这些QTL中的80个中，基于其效果和功能注释的重要性，突出显示了855 CV。随后在MON，NOR和HOL奶牛上进行的GWAS验证了8、9和23个QTL的生产性状。SCS为0、0和1；乳房形态特征分别为4、1、8。在确认的54个QTL中，有47个在公牛中鉴定出的CV比标准50K芯片的单核苷酸多态性（SNP）具有更显着的影响。每个经过验证的QTL的最佳CV位于与生产（36 QTL）或乳房（9 QTL）性状在功能上相关的基因中。使用这种分两步的GWAS方法，我们鉴定并验证了54个QTL，其中QC主要位于功能候选基因内，并解释了具有重要经济意义的乳品性状的遗传变异的6.3％（乳房性状）和37％（生产性状）。这些CV现在包含在用于评估法国奶牛的芯片中，并且可以集成到常规基因组评估中。经济上重要的乳品性状的遗传变异的3％（乳房性状）和37％（生产性状）。这些CV现在包含在用于评估法国奶牛的芯片中，并且可以集成到常规基因组评估中。经济上重要的乳品性状的遗传变异的3％（乳房性状）和37％（生产性状）。这些CV现在包含在用于评估法国奶牛的芯片中，并且可以集成到常规基因组评估中。