Developing an integrative approach to phenomics for industrial, biomedical and environmental applications

用于高通量和自动测量分子水平现象的组学技术在生物学的所有领域都取得了重大进展。这种理解生物学的还原论方法是富有成效的，但它们在理解我们简单地称之为表型的复杂、多维综合实体方面的能力有限。一种突破性的描述和量化表型的新方法是表型组学-在整个生物体范围内获取高维数据。表型组学的挑战是相当大的，因为表型组的信息含量和复杂性使基因组相形见绌。然而，在医学和植物科学中，表型组学方法的发展取得了重大进展，这表明它有可能改变我们对生物现象的解释和理解。该奖学金将推动新一代水生胚胎表型组学技术的发展，并验证其在解决一些最紧迫的生物医学、环境和工业问题和挑战方面的应用。水生胚胎体积小，灵敏度高，时间、空间和功能变化率高，是表型组学研究的理想模型。我的奖学金将以胚胎表型学（EP）为基础，这是我在过去八年中开发的一项尖端技术，能够以前所未有的时间，空间和功能分辨率全自动同时高通量筛选数百种不同水生胚胎的表型（Tills等人2018 PLoS Biology， 16: e30000074）。EP由两种协同技术组成：定制高通量生物成像硬件（HEIF资助）和自动化分析软件（NERC资助）。EP平台克服了研究生物体生命史最动态时期的挑战，通过对数百万张图像的全自动采集和分析，同时对数百个发育中的水生胚胎进行分析，它已被证明在测量不同物种对长期和短期暴露于环境驱动因素的致命和亚致命生物反应方面非常有效。以EP为基础，我将更广泛地使用人工智能（AI）和机器学习来训练模型：(I)从水生胚胎视频中识别和量化表型性状；（ii）整合(i)中产生的每个胚胎的数千个特征，以识别功能组合现象信号。目前EP的硬件部件为定制件和定制高端消费件，成本高，创新障碍大。这项奖学金将支持我开发LabEP和FieldEP，这两种具有变革性和可访问性的表型组学技术。LabEP将是现有EP技术的优化版本，具有增强的功能和更低的成本。定制3D打印部件的开发，用于同时高通量筛选数千个胚胎的微流控培养设备，以及将用于药物化合物筛选和生态毒理学验证的强大AI模型。FieldEP将是一个长期的现场可部署的摄像系统，它将使用定制的训练过的人工智能模型与环境传感器协同作用，自动量化现场的生物反应。FieldEP将在普利茅斯（英国第一个拟议的海洋公园）和南极洲的物流挑战环境中，对海洋环境中混合和波动环境驱动因素的现象响应进行量化验证。这项奖学金将支持我通过实验室（高通量）和现场（混合和波动的环境驱动因素）新兴技术的前沿方法创新，开发一种综合的表型组学方法，并将其应用于生物学中一些最紧迫的挑战。

Omics technologies for high-throughput and automated measurement of phenomena at the molecular level have enabled significant advances in all areas of biology. Such reductionist approaches to understanding biology have been productive, but they are limited in their capacity to provide understanding of the complex, multidimensional integrated entity we refer to simply to as the phenotype. A ground-breaking new approach to describing and quantifying the phenotype is phenomics - the acquisition of high-dimensional data on an organism-wide scale. The challenge of phenomics is considerable due to the information content and complexity of the phenome dwarfing that of the genome. Yet significant advances in the development of phenomics approaches in medicine and plant science are demonstrating its potential to transform our interpretation and understanding of biological phenomena. This fellowship will drive the development of a new generation of technologies for phenomics in aquatic embryos and validate their application for addressing some of the most pressing biomedical, environmental and industrial questions and challenges. The small size, high sensitivity and high rates of temporal, spatial and functional change make aquatic embryos excellent models for phenomics. My fellowship will be underpinned by EmbryoPhenomics (EP), a cutting-edge technology that I have developed over the past eight years that enables the fully automated simultaneous high-throughput screening of the phenome of hundreds of different aquatic embryos with unprecedented temporal, spatial and functional resolution (Tills et al 2018 PLoS Biology, 16: e30000074). EP consists of two synergistic technologies: custom high-throughput bioimaging hardware (HEIF funded) and automated analytical software (NERC funded). The EP platform surmounts the challenge of studying the most dynamic period of an organism's life history via the fully automated acquisition and analysis of millions of images, of hundreds of developing aquatic embryos simultaneously and it has proven highly effective at the measurement of lethal and sublethal biological responses to long- and short-term exposures to environmental drivers in different species.With EP as a foundation, I will use artificial intelligence (AI), and machine learning more broadly, to train models for: (i) the identification and quantification of phenotypic traits from video of aquatic embryos; and (ii) the integration of the thousands of traits for each embryo produced in (i) to identify functional combinatorial phenomic signals. Currently the hardware components of EP are bespoke parts and custom high-end consumer parts, with a high cost and barrier to innovation. This fellowship will support me in developing LabEP and FieldEP - two transformative and accessible technologies for phenomics. LabEP will be an optimised version of the existing EP technology with enhanced functionality and a lower cost drawing on; the development of custom 3D printed parts, microfluidic culture devices for high-throughput screening of thousands of embryos simultaneously and powerful AI models that will be validated for both pharmaceutical compound screening and ecotoxicology. FieldEP will be a long-term field-deployable camera system that will automatically quantify biological responses in situ using custom trained AI models in synergy with environmental sensors. FieldEP will be validated for quantifying phenomic responses to mixed and fluctuating environmental drivers in marine environments in both Plymouth, the UK's first proposed marine park, and the logistically challenging environment of Antarctica. This fellowship will support me in developing an integrative approach to phenomics via the innovation of cutting-edge approaches enabled by emerging technologies for the laboratory (high-throughput) and field (mixed and fluctuating environmental drivers) and their application to some of the most pressing challenges in biology.