Developing insect population models to support the design of GM control strategies

马蝇是一种全球分布的主要公害昆虫，能够机械传播一系列重要的牲畜和人类病原体。据估计，其咬伤活动造成的压力和伤害每年给美国养牛业造成约10亿美元的损失。由于最近甘蔗和其他蔬菜废弃物管理方面的变化，预计其在巴西和澳大利亚等地区的影响将会增加。因此，更好的控制稳定苍蝇的方法将有利于动物福利、生物能源生产和食品生产。该项目的工业东道主Oxitec在开发和释放转基因菌株来控制昆虫种群方面处于世界领先地位。对于蚊子来说，这通常涉及释放非常大量的改良雄性昆虫，以击败野生型雄性蚊子。这影响很小，因为雄性蚊子不吸血。为了使用类似的方法来控制稳定的苍蝇种群，更需要工具来设计最佳有效的释放策略，因为雄性和雌性稳定的苍蝇都是以血液为食的。拟议的项目将开发基于过程的稳定苍蝇种群模型，并使用现代统计方法将它们与数据进行匹配。基于过程的模型明确地模拟了出生率和死亡率等生物过程，而不是简单地模拟丰富度。它们相对复杂，在历史上很难适应，但它们更适合于探索人口对气候变化或控制战略等异常情况的反应。最近，已经发展了更好的统计方法来将复杂的模型与数据进行拟合，如近似贝叶斯计算(ABC)。皮尔布赖特的数学生物学小组有成功地使用尖端统计方法将复杂的生物模型与数据拟合的历史。在最近北欧爆发Schmallenberg病毒期间，该小组使用近似贝叶斯计算将复杂的疾病传播模型与疫情爆发的早期阶段进行匹配，从而能够对提供给欧盟委员会以做出疫情管理决策的关键疾病传播参数做出推断。该模型的结论后来得到了实验验证。该小组也有促进接受政策决定的模拟产出的历史。例如，该小组提供的疾病传播模拟最近被用来帮助为一种新的蓝舌病疫苗产品提供许可，在2007年英国BTV爆发期间，该小组提供了实时模拟输出，以回应政策制定者的询问。基于过程的昆虫种群建模是该小组研究媒介传播疾病传播的合乎逻辑的下一步，因为它将使气候变化和新的控制策略对昆虫媒介种群的影响得到更准确的预测。出于几个原因，稳定的苍蝇是一种理想的物种。首先，如上所述，该物种在世界许多地区具有重大的和日益增长的直接影响，代表着一个巨大的潜在市场。其次，有机会在开发转基因控制产品的同时开发该模型，然后使用项目输出来设计新菌株的最佳释放策略，以支持其吸收。这位学术合作伙伴维护着英国唯一一群稳定的苍蝇，并能够通过材料和技术支持该项目。EMBRAPA的同事们最近开始收集他们愿意为该项目分享的人口观测的大型数据集，并且在之前的研究活动中已经开发了一个概要的人口模型。

The stable fly is a major nuisance insect with a global distribution and is capable of mechanically transmitting a range of important livestock and human pathogens. The stress and injury caused by its biting activity are estimated to cost the US cattle industry around $1billion/year. Its impact is projected to increase in regions including Brazil and Australia as a consequence of recent changes to the management of sugar cane and other vegetable waste. Better control methods for stable flies would therefore benefit animal welfare, bioenergy production and food production.The industrial host of this project, Oxitec, is a world leader in developing methods for manipulating insect populations through the development and release of genetically modified strains. For mosquitoes, this typically involves the release of very large numbers of modified male insects to outcompete wild-type males. This has minimal impact because male mosquitoes do not blood-feed. To use similar approaches to control stable fly populations, there is a greater need for tools to design optimally efficient release strategies because male and female stable flies both blood-feed.The proposed project will develop process-based models of stable fly populations and use modern statistical approaches to fit them to data. Process-based models explicitly model biological processes such as birth and death rates rather than simply modelling abundance. They are relatively complex and the difficulty of fitting them has historically been a limitation, but they are better for exploring the responses of populations to unusual circumstances such as climate change or control strategies. Recently, better statistical approaches to fitting complex models to data have been developed, such as Approximate Bayesian Computation (ABC).The Mathematical Biology group at Pirbright has a history of successfully using cutting-edge statistical approaches to fit complex biological models to data. During the recent outbreak of Schmallenberg virus in northern Europe, the group used Approximate Bayesian Computation to fit a complex disease transmission model to the early stages of the outbreak, allowing inferences to be made about key disease transmission parameters which were provided to the European Commission to make outbreak management decisions. The conclusions of the model were later validated experimentally. The group also has a history of facilitating the acceptance of modelling outputs for policy decisions. For example, disease spread simulations provided by the group were recently used to help make the case for licensing a novel bluetongue vaccine product, and during the BTV outbreak in the UK in 2007 the group provided simulation outputs in real-time in response to queries from policymakers.Process-based insect population modelling is a logical next step for the group's research into the spread of vector-borne diseases, as it will allow the effects of climate change and novel control strategies on insect vector populations to be predicted with greater certainty. The stable fly is an ideal species to begin with for several reasons. Firstly, as outlined above, the species is associated with a substantial and growing direct impact in many areas of the world and represents a large potential market. Secondly, an opportunity exists to develop the model in parallel with a GM control product and then use the project outputs to design optimal release strategies of the new strains, supporting its uptake. The academic partner maintains the only colony of stable flies in the UK and is able to support the project via materials and know-how. Colleagues at EMBRAPA have recently begun collecting a large dataset of population observations that they are willing to share for the project, and an outline population model was already developed during previous research activities.