DMS/NIGMS 2: Spatial, Multi-Host Petri Net Models for Zoonotic Disease Forecasting

野生动物传播的疾病在全球范围内对人类健康构成越来越大的威胁，但 预测人畜共患病原体如何传播仍然是一个重大挑战。人畜共患病是复杂的， 其行为受社会、生态、遗传、 和进化因素。了解生物和非生物因素对准确的疾病预测的贡献，是管理快速环境变化的新风险和改进复杂生态系统的机制模型的当务之急。监测人畜共患病病原体及其寄主物种的实地研究通常认为，观察到宿主的高流行率是强有力的证据，表明(I)宿主是病原体的‘储存库’，将其维持在稳定的种群水平；以及(Ii)宿主是向其他物种溢出的持久来源。然而，生态模型已经表明，水库的状况可能强烈地依赖于背景，受到外部因素的调节，包括与其他物种的相互作用和栖息地的碎片化。因此，需要一种结合复杂系统建模、数据科学和风险分析的跨学科方法来对人畜共患病溢出动力学进行建模。该项目将在亚利桑那州凤凰城地区建立新的大规模、多宿主的汉坦病毒肺综合征(HPS)和谷热(球孢子菌病)疾病的机制模型，以调查宿主作为疾病宿主的稳定性和上下文敏感性，并确定改善预测和确定旨在减少疾病负担的有效干预措施所需的关键未来数据源。这两种疾病都是美国西南部的地方病，据信是由啮齿动物宿主传播的，但它们由不同类型的病原体(分别是病毒和真菌)引起，并显示出不同的病例趋势。该项目将使用模块化的、可扩展的、易于可视化为机械网络图的Petri网模型，这使它们成为探索添加或移除主机和改变土地使用预期如何改变疾病动态的宝贵工具。 在培训和外展方面，该项目将实施和评估以下举措：(I)在公共外展活动中交流成果；(Ii)建立以课程为基础的本科生研究经验(CURE)，侧重于疾病建模；以及(Iii)建设研究山谷热和减轻疫情的能力。在推广方面，该项目将创建一个免费的网络应用程序，供公众与模块化疾病模型及其可视输出进行互动。除了展示HPS和Valley Fever的项目结果外，该Web应用程序还将链接到介绍该项目将开发的PETRI网的互动教科书。对于治愈课程，学生将专注于合成人畜共患病数据以改进风险建模，为亚利桑那州立大学约7,000名面对面和在线生物学专业的学生提供急需的研究机会。最后，该项目将为学术和公共卫生研究人员组织一次能力建设讲习班，向他们介绍Petri网资源，并确定改进预测所需的数据。

Diseases that spillover from wild animals pose an increasing threat to human health worldwide, but forecasting how zoonotic pathogens spread remains a major challenge. Zoonotic diseases are complex, spatially and temporally evolving systems whose behaviors are influenced by social, ecological, genetic, and evolutionary factors. Understanding the contributions of biotic and abiotic factors to accurate disease forecasting is an urgent priority for managing the emerging risks of rapid environmental change and for improving mechanistic models of complex ecological systems. Field studies monitoring zoonotic pathogens and their host species have typically assumed that observing high host prevalence is strong evidence that (i) the host is a ‘reservoir’ of the pathogen, maintaining it at a stable population level; and (ii) the host is a persistent source of spillover into other species. However, ecological models have shown that reservoir status can be strongly context dependent, mediated by extrinsic factors including interactions with other species and habitat fragmentation. An interdisciplinary approach combining complex systems modeling, data science, and risk analysis is therefore needed to model zoonotic spillover dynamics. This project will construct novel largescale, multi-host mechanistic models of the Hantavirus Pulmonary Syndrome (HPS) and Valley Fever (coccidioidomycosis) diseases in the Phoenix, Arizona metro area to investigate the stability versus context-sensitivity of hosts as disease reservoirs and identify key future data sources required to improve forecasting and identify effective interventions aimed at reducing disease burden. Both diseases are endemic to the southwestern United States and are believed to be spread by rodent hosts, but they are caused by different types of pathogens (viruses and fungi, respectively) and show divergent case trends. The project will use Petri Net models, which are modular, scalable, and readily visualized as mechanistic network diagrams, which makes them a valuable tool for exploring how adding or removing hosts and changing land use are expected to change disease dynamics. For training and outreach, the project will implement and evaluate initiatives to (i) communicate results in public outreach events; (ii) construct a course-based undergraduate research experience (CURE) focused on disease modeling; and (iii) build capacity for researching Valley Fever and mitigating outbreaks. For outreach, the project will create a free web app for public interaction with modular disease models and their visual outputs. In addition to presenting project results for HPS and Valley Fever, the web app will be linked to an interactive textbook introducing Petri Nets to be developed by the project. For the CURE class, students will focus on synthesizing zoonotic disease data to improve risk modeling, providing urgently needed research opportunities for ASU’s approximately 7,000 in-person and online biology majors. Lastly, the project will organize a capacity-building workshop of academic and public health researchers to introduce them to Petri Net resources and to identify data needs for improved forecasting.