PIPP Phase I: Predicting and Preventing Epidemic to Pandemic Transitions

PIPP 第一阶段：预测和预防流行病向大流行病的转变

• 批准号: 2200052
• 负责人: Ioannis Paschalidis
• 金额: $ 100万
• 依托单位: Trustees of Boston University
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2022
• 资助国家: 美国
• 起止时间: 2022-08-01 至 2025-01-31
• 项目状态: 未结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=2200052&HistoricalAwards=false
• 关键词: PIPP; Phase; Predicting; Preventing; Epidemic

The COVID-19 pandemic and its effects, both in terms of the millions of lives lost and the trillions in estimated costs, are a recent example of the devastation pandemics can cause. Any discernible progress in the prediction, early detection, and rapid response would have significant impacts on human welfare. The overarching goal of this project is to develop a comprehensive strategy and the required science for predicting and preventing future pandemics. Predicting a pandemic at its pre-emergence, zoonotic stage requires considering millions of undescribed viruses thought to exist in mammals and birds, which could lead to many false alarms. On the other hand, detecting a pandemic after it has spread widely is too late. Instead, this project will develop methods for detecting when an emerging pathogen has spilled over from its natural animal reservoir into humans, causing a small, localized disease cluster, and will seek to develop a suite of rapid response and mitigation strategies. The research agenda will form the basis of a research Center to undertake a longer-term effort. This project offers educational and training opportunities for graduate students and post-docs. The research is organized around four tasks that map to a natural progression from prediction and detection to prevention. Task 1 seeks to identify location hotspots of pathogen emergence and to compile ranked lists of the most likely zoonotic pathogens that could cause an initial outbreak. Task 2 will focus on detecting disease anomalies in healthcare settings with methods applicable to resource-limited settings, leveraging alternative data sources from social media, web search, cell phone mobility patterns, local case reports, and death reports. Task 3 will consider the more detailed characterization of a pathogen causing a local disease cluster. It will also develop network-based disease spread models to predict if, and under what conditions, the local disease cluster is likely to evolve into a pandemic. Task 4 will focus on mitigation and response strategies, including individual therapeutics and vaccines, issues of global governance, and decision-making tools to deploy control mechanisms in the form of travel restrictions, lockdowns, social distancing and mask-wearing directives, and drug/vaccine resource allocation. To evaluate the developed framework, the team will apply it to recent historical epidemics and pandemics, considering COVID-19, H1N1, and Ebola. The research team spans a large multidisciplinary space, including biology, ecology, epidemiology, medicine (infectious diseases, virology and microbiology), computer & information science & engineering, and social sciences (behavioral sciences, health policy, and emerging media).This award is supported by the cross-directorate Predictive Intelligence for Pandemic Prevention Phase I (PIPP) program, which is jointly funded by the Directorates for Biological Sciences (BIO), Computer Information Science and Engineering (CISE), Social, Behavioral and Economic Sciences (SBE) and Engineering (ENG).This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

COVID-19大流行及其影响，无论是数百万人丧生还是数万亿美元的估计成本，都是大流行病可能造成破坏的最新例子。在预测、早期发现和快速反应方面的任何明显进展都将对人类福祉产生重大影响。 该项目的总体目标是制定一项全面战略和预测和预防未来大流行病所需的科学。在出现之前预测大流行，人畜共患病阶段需要考虑数百万被认为存在于哺乳动物和鸟类中的未描述病毒，这可能导致许多错误警报。另一方面，在大流行病广泛传播后才发现为时已晚。相反，该项目将开发用于检测新出现的病原体何时从其天然动物宿主溢出到人类身上的方法，并将寻求开发一套快速反应和缓解策略。研究议程将构成研究中心开展长期工作的基础。该项目为研究生和博士后提供教育和培训机会。该研究围绕四项任务组织，这些任务映射到从预测和检测到预防的自然进展。任务1旨在确定病原体出现的地点热点，并编制最有可能导致初始暴发的人畜共患病病原体的排名列表。任务2将侧重于使用适用于资源有限环境的方法检测医疗保健环境中的疾病异常，利用来自社交媒体，网络搜索，手机移动模式，本地病例报告和死亡报告的替代数据源。任务3将考虑导致当地疾病群的病原体的更详细的特征。它还将开发基于网络的疾病传播模型，以预测当地疾病集群是否以及在何种条件下可能演变为大流行病。任务4将侧重于缓解和应对策略，包括个体疗法和疫苗、全球治理问题以及决策工具，以部署旅行限制、封锁、社交距离和戴口罩指令等形式的控制机制，以及药物/疫苗资源分配。为了评估开发的框架，该团队将把它应用于最近的历史流行病和大流行病，考虑COVID-19，H1N1和埃博拉病毒。研究团队跨越了一个庞大的多学科空间，包括生物学、生态学、流行病学、医学（传染病、病毒学和微生物学）、计算机&信息科学&工程和社会科学（行为科学，卫生政策和新兴媒体）。该奖项由跨部门的大流行病预防第一阶段预测情报（PIPP）计划支持，这是由生物科学（BIO），计算机信息科学与工程（CISE），社会，行为和经济科学（SBE）和工程（ENG）该奖项反映了NSF的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

项目成果

Convergence of Actor-Critic with Multi-Layer Neural Networks

Actor-Critic 与多层神经网络的融合

• 发表时间: 2023
• 期刊: Advances in neural information processing systems
• 作者: Tian, H.;Olshevsky, A.;Paschalidis, I.C.
• 通讯作者: Paschalidis, I.C.

Site-Wide HPC Data Center Demand Response

全站 HPC 数据中心需求响应

• DOI: 10.1109/hpec55821.2022.9926322
• 发表时间: 2022
• 期刊: IEEE High Performance Extreme Computing Conference
• 作者: Wilson, Daniel C.;Paschalidis, Ioannis Ch.;Coskun, Ayse K.
• 通讯作者: Coskun, Ayse K.

Improved prediction of MHC-peptide binding using protein language models.

• DOI: 10.3389/fbinf.2023.1207380
• 发表时间: 2023
• 期刊: FRONTIERS IN BIOINFORMATICS
• 作者: Hashemi, Nasser;Hao, Boran;Ignatov, Mikhail;Paschalidis, Ioannis Ch;Vakili, Pirooz;Vajda, Sandor;Kozakov, Dima
• 通讯作者: Kozakov, Dima

Threatening the Future of Global Health — NIH Policy Changes on International Research Collaborations

威胁全球健康的未来 — NIH 国际研究合作政策变化

• DOI: 10.1056/nejmp2307543
• 发表时间: 2023
• 期刊: New England Journal of Medicine
• 影响因子: 158.5
• 作者: Ko, Albert I.;Karim, Salim S.;Morel, Carlos;Swaminathan, Soumya;Daszak, Peter;Keusch, Gerald T.
• 通讯作者: Keusch, Gerald T.

Distributionally Robust Multiclass Classification and Applications in Deep Image Classifiers

分布式鲁棒多类分类及其在深度图像分类器中的应用

• DOI: 10.1109/icassp49357.2023.10095775
• 发表时间: 2023
• 期刊: and Signal Processing (ICASSP
• 作者: Chen, Ruidi;Hao, Boran;Paschalidis, Ioannis Ch.
• 通讯作者: Paschalidis, Ioannis Ch.