Design of breeding programs to improve honeybee health and production

这个博士项目将设计育种计划，通过研究人员，从业者，养蜂人和养蜂人之间的独特合作来改善蜜蜂的健康和生产。蜜蜂由于害虫和环境变化而受到压力。关于哪种蜜蜂最适合不同的环境条件和管理方法，存在着不同的观点。这些观点主要基于蜜蜂的形态和生产力特征，定量证据和分析有限。这个问题可以通过三管齐下的方法来解决：（i）收集不同类型蜜蜂、条件和实践的定量证据，（ii）将表型变异与所述证据相关联，以及（iii）设计现代，高效和可持续的数据驱动育种计划。为此，该项目将分为三个工作包。首先，将从英国和/或新西兰的殖民地样本中收集定量证据。焦点表型性状将是健康状况和生产沿着最先进的高频率和高容量物联网（IoT）群体传感器和元数据（管理，位置和天气）。此外，还将对菌落进行全基因组测序。具体而言，回归表型蜜蜂形态，全基因组序列变异，管理，位置和天气描述符。基因组和位置将通过核算系统发育和空间关系来建模，以解开遗传和环境变异。此外，将使用机器学习评估与物联网殖民地数据的关联，以建立健康状况的预测模型。这些结果将为遗传、管理和环境影响的差异提供定量证据。第三，前两个工作包的结果将用于建立计算机蜜蜂育种框架。该框架将用于设计最适合不同条件和实践的育种计划。将特别注意空间分布，因此不可避免的杂交不同的蜜蜂类型。此外，重点将放在设计可持续的育种计划（在遗传多样性方面），可以解决当前和未来的挑战。总之，结果将是一套实用的育种计划，将改善健康和生产在不同的条件和做法。

This PhD project will design breeding programs to improve honeybee health and production through a unique collaboration between researchers, practitioners, beekeepers and bee-farmers.Honeybees are under pressure due to pests and environmental changes. There are competing views on what type of a honeybee are best suited different environmental conditions and management practices. These views are largely based on bee morphology and productivity traits, with limited quantitative evidence and analysis.The issue can be addressed via a three-pronged approach of (i) collecting quantitative evidence for different types of bees, conditions and practices, (ii) associating phenotypic variation to said evidence and (iii) designing modern, efficient and sustainable data-driven breeding programs.To this end the project will be organised in three work-packages.First, quantitative evidence will be collected from a sample of colonies in the UK and/or New Zealand. The focal phenotype traits will be health status and production along with the state-of-the-art high-frequency and high-volume Internet-of-things (IoT) colony sensors and meta-data (management, location, and weather). In addition, colonies will be whole-genome sequenced.Second, the collected data will be analysed. Specifically, regressing phenotypes on bee morphology, whole-genome sequence variants, management, location and weather descriptors. The genome and location will be modelled by accounting phylogenetic and spatial relationships to disentangle genetic and environmental variation. In addition, associations with IoT colony data will be evaluated with machine learning to build predictive models for health status. These results will provide quantitative evidence about differences due to genetic, management and environmental effects.Third, the results from the first two work-packages will be used to build an in-silico honeybee breeding framework. This framework will be used to engineer most suitable breeding programs for different conditions and practices. Particular attention will be given to spatial distribution and therefore unavoidable hybridization between different bee types. Further, emphasis will be given to design sustainable breeding programs (in terms of genetic diversity) that can address current and future challenges. In summary the outcome will be a set of practical breeding schemes that will improve health and production under different conditions and practices.