Inferring genetic associations among dairy cow behavior components using biomarkers, genetic marker, genome sequence and technical data

奶牛行为是影响奶牛群盈利能力的主要因素，也是动物福利和疾病的指标。行为是一个复杂的网络行为模式，以响应环境和社会刺激，以及人类的处理。农业技术的进步导致了世界范围内奶牛饲养系统的变化。逐渐增加的畜群规模，更少的时间来照顾动物和安装现代技术，如自动挤奶系统（AMS）意味着有限的人与牛的互动。反之亦然，改善奶牛对技术环境的行为反应（奶牛-AMS相互作用）对于高效生产和福利友好条件变得越来越重要，并有助于简化“奶牛处理”和减少劳动时间。AMS产生客观的行为特征（即，“大数据”）与可加工性、可挤奶性和健康相关联，其可以被实施到基因组选择工具中。然而，奶牛学习和社会行为基因组学的研究还不够深入，这从根本上影响着奶牛群的管理、生产和福利。此外，选择适应AMS环境的行为需要深入了解行为性状之间的遗传关联，以及行为性状与生产和健康性状之间的遗传关联，这是目前未知的。在过去的十年中，来自JLU和MLU两个研究小组的遗传学家和兽医在德国实施了所谓的奶牛训练集，用于基于密集表型和基因型荷斯坦黑白花奶牛的基因组选择。包括超过20'000头基因分型奶牛的这种综合数据集是实施健康性状基因组选择的基础。数据库包括几个使用AMS的大规模牛群，能够基于5 '000头基因分型奶牛对AMS行为性状进行基因组分析，这些奶牛被输入序列数据。AMS技术纵向行为数据将与其他行为类别（喂养和活动行为，社会行为，生殖行为，对人类处理的行为反应）和行为生物标志物（心率，皮质醇浓度，直肠温度）合并，以便通过全基因组多效性方法和全基因组关联研究推断奶牛行为的基因组机制。结果将与功能基因数据库进行注释，目的是确定候选基因和生物学途径，有助于不同的牛行为类别。已确定的遗传变异将在增强的遗传评估中考虑，旨在开发奶牛行为的选择工具。

Cow behavior is a major factor influencing dairy herd profitability and is an indicator for animal welfare and disease. Behavior is a complex network of behavioral patterns in response to environmental and social stimuli, and to human handlings. Advances in agriculture technology led to changes in dairy cow husbandry systems worldwide. Gradually increasing herd sizes, less time availability to take care for the animals and installed modern technology such as automatic milking systems (AMS) imply limited human-cow interactions. Vice versa, improving cow behavior response to technical environment (Cow-AMS interactions) becomes increasingly important for efficient production and welfare friendly conditions, and contributes to simplified “cow handling” and reduced labour time. AMS generate objective behavior traits (i.e., “big data”) linked to workability, milkability and health, which can be implemented into genomic selection tools. However, there is insufficient understanding of learning and social behavioral genomics in cows, which affect management, production and welfare in dairy herds essentially. Moreover, selection for adapted behavior to AMS environments requires deep knowledge of the genetic associations among behavior traits, and between behavior traits with production and health traits, which are currently unknown. During the past decade, geneticists and veterinarians from both research groups JLU and MLU implemented so-called cow training sets for genomic selection based on dense phenotyped and genotyped Holstein Friesian cows in Germany. Such comprehensive dataset including more than 20’000 genotyped cows was the basis for implementing genomic selection for health traits. The data pool comprises several large-scale herds using AMS, enabling genomic analyses for AMS behavior traits based on 5’000 genotyped cows imputed to sequence data. AMS technical longitudinal behavior data will be merged with other behavior categories (feeding and activity behavior, social behavior, reproduction behavior, behavior response to human handlings) and behavior biomarkers (heart rate, cortisol concentrations, rectal temperature) in order to infer genomic mechanisms of cow behavior via genome-wide pleiotropy approaches and genome-wide association studies. Results will be annotated with functional gene databases, aiming on the identification of candidate genes and biological pathways contributing to the different cattle behavior categories. Identified genetic variants will be considered in enhanced genetic evaluations, aiming on selection tool developments for cow behaviour.