CICI:UCSS:Improving the Privacy and Security of Data for Wastewater-based Epidemiology

基于废水的流行病学(WBE)领域收集和分析来自污水处理系统的与公共健康相关的数据。这些数据包括身份、种族、行为、消费、污染和病原体感染的遗传和化学生物标记物。最近，随着研究人员转向废水样本来估计当地人口中SARS-CoV-2的感染水平，并为应对大流行的公共卫生对策提供信息，对WBE的兴趣激增。由于从废水收集的数据是在人口层面上收集的，它们通常被认为是匿名的，因此不受健康隐私或其他监管保护。这些数据在共享时最有用，但共享会给个人以及社区、学校和政府带来安全和隐私问题。该项目正在开发技术和工具，以支持WBE数据集的合法使用，同时将滥用降至最低，并正在设计保护个人和组织隐私的协议。这些工具将通过使组织能够保护其数据不被恶意使用来鼓励其广泛参与WBE联盟，这反过来将有助于改善普通公众的健康并减少发病率和死亡率。该项目解决了三个关键的安全威胁：个人查询构成的威胁；共享构成的威胁，无论是通过联合查询还是大规模关联研究；以及废水样本中存在人类遗传物质构成的威胁。这些威胁通过以下方式得到缓解：(1)使用同态加密支持对废旧数据的安全查询；(2)基于安全计算的联合数据分析；(3)通过差分隐私保护基因组隐私。正在实施这些组成部分，以使不同的实体能够执行查询处理并训练关于联合数据库的机器学习模型，同时保持每个数据库的私密性和格式。该系统正在使用与重要威胁场景匹配的使用案例的真实世界数据集进行评估。该奖项反映了NSF的法定使命，并通过使用基金会的智力优势和更广泛的影响审查标准进行评估，被认为值得支持。

The field of Wastewater-Based Epidemiology (WBE) collects and analyzes data from sewage systems that relate to public health. These data include genetic and chemical biomarkers of identity, ethnicity, behavior, consumption, pollution, and pathogenic infections. Interest in WBE exploded recently, as researchers turned to wastewater samples to estimate SARS-CoV-2 infection levels in local populations and inform public health responses to the pandemic. Because data collected from wastewater are aggregated at the population level, they are typically assumed to be anonymous, and are therefore not subject to health privacy or other regulatory protections. These data are most useful when shared, yet sharing raises security and privacy issues, for individuals as well as neighborhoods, schools and governments. The project is developing technology and tools to support legitimate uses of WBE datasets while minimizing abuse, and it is designing protocols for protecting individual and organizational privacy. These tools will encourage wide participation in WBE consortia by enabling organizations to protect their data from adverse uses, which in turn will help improve the health of the general public and reduce morbidity and mortality.The project addresses three key security threats: those posed by individual queries; those posed by sharing, whether through joint queries or large-scale association studies; and those posed by the presence of human genetic material in wastewater samples. These threats are mitigated by: (1) supporting secure queries on wastewaster data using homomorphic encryption; (2) enabling federated data analysis based on secure computation; and (3) protecting genomic privacy with differential privacy. These components are being implemented to allow different entities to execute query processing and train Machine Learning models on joint databases while maintaining the privacy and format of each database. The system is being evaluated using real-world datasets with use cases that match important threat scenarios.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.