Using comparative genomics to develop 'digital twins' to support SMART ecotoxicological predictions

我们快速变化的世界正使关键的生态系统处于前所未有的环境压力之下，这种压力包括暴露于各种化学毒物中。本博士的总体目标是利用基因组和其他性状资源来提供对有毒物种敏感性的机械信息估计。有效保护生态系统需要化学毒性知识。然而，这些信息必须在3r目标的背景下获得，即努力减少、改进和取代动物的使用。在没有化学接触实验的情况下准确预测物种对有毒物质的敏感性，将是实现这一目标的重要一步。然而，实现这一目标需要对相关的生物学途径、它们在物种间的保护以及便于比较和评估的框架有深刻的机械理解。幸运的是，这样的目标现在是可以实现的，因为快速增长的基因组资源包含了控制污染物敏感性的分子成分的比较数据宝库。这项博士研究将对了解化学物质对生态系统的影响做出宝贵的贡献，而无需进行动物接触。博士生将获得必要的技能和理解，以生成有机体的硅表征（称为“数字双胞胎”），这将为比较生态毒理学评估提供信息。为了产生这样的“数字双胞胎”，博士生将研究和设计一个系统，将分子信息、物种特征数据和建模工具集成在一个模块化框架内。这样做的目的是产生一个包含各种数据类型（例如基因组数据、能量和表型特征）的自动化数据基础设施，可用于快速检索与毒物敏感性预测相关的跨物种信息。一旦基础设施建立起来，博士生将努力开发方法（例如使用人工智能）来帮助更好地预测各种物种特征对敏感性的复杂和多变量贡献。这种人工智能方法的输出将与已建立的生态毒理学数据相结合，以验证物种特征和观察到的敏感性之间的联系。自动化的数据基础设施及其产生的机制洞察力将提高生态毒理学家和环境监管机构预测未测试物种敏感性的能力

Our rapidly changing world is placing critical ecosystems under unprecedentedenvironmental pressures, pressure that includes exposure to a wide-range of chemicaltoxicants. The overarching aim of this PhD is to harness genomic and other traitresources to deliver mechanistically informed estimates of toxicant species sensitivity.The efficient protection of ecosystems requires knowledge of chemical toxicity. However,such information has to be obtained within the context of the 3Rs goal, i.e. the effort toReduce, Refine and Replace the use of animals. Accurate prediction of species sensitivityto toxicants without chemical exposure experiments would represent a major step towardfulfilling this ambition. However, realising this aim requires a deep mechanisticunderstanding of relevant biological pathways, their conservation across species, and aframework to facilitate easy comparison and assessment. Fortunately, such objectivesare now achievable, as rapidly increasing genomic resources contain a treasure trove ofcomparative data on the molecular components governing pollutant sensitivity. This PhDresearch will make an invaluable contribution to understanding chemical effects onecosystems without performing animal exposures.The PhD student will obtain the skills and understanding necessary to generate in silicorepresentations of organisms (termed 'digital twins') that will inform comparativeecotoxicological assessment. To generate such 'digital twins', the PhD student willinvestigate and design a system to integrate molecular information, species trait data andmodelling tools within a modular framework. This will be done with the aim of producingan automated data infrastructure containing varied data types (e.g. genomic data,energetic and phenotypic traits), that can be used to rapidly retrieve cross-speciesinformation relevant to toxicant sensitivity predictions. Once the infrastructure isestablished, the PhDstudent will endeavor to develop approaches (e.g. using artificialintelligence) to help better predict the complex and multi-variate contributions made byvarious species characteristics to sensitivity. The output of such artificial intelligenceapproaches will be combined with established ecotoxicological data to validate linksbetween species characteristics and observed sensitivity. Both the automated datainfrastructure and the mechanistic insight it produces will increase the capacity ofecotoxicologists and environmental regulators to predict sensitivity in untested species