Institute of Environmental Science and Department of Environmental Science, Hankuk University of Foreign Studies, Yong-In, Korea.
PLoS One. 2021 Apr 2;16(4):e0249528. doi: 10.1371/journal.pone.0249528. eCollection 2021.
Recognizing that microbial community composition within the human microbiome is associated with the physiological state of the host has sparked a large number of human microbiome association studies (HMAS). With the increasing size of publicly available HMAS data, the privacy risk is also increasing because HMAS metadata could contain sensitive private information. I demonstrate that a simple test statistic based on the taxonomic profiles of an individual's microbiome along with summary statistics of HMAS data can reveal the membership of the individual's microbiome in an HMAS sample. In particular, species-level taxonomic data obtained from small-scale HMAS can be highly vulnerable to privacy risk. Minimal guidelines for HMAS data privacy are suggested, and an assessment of HMAS privacy risk using the simulation method proposed is recommended at the time of study design.
认识到人类微生物组内的微生物群落组成与宿主的生理状态有关，这引发了大量的人类微生物组关联研究（HMAS）。随着可公开获得的 HMAS 数据的不断增加，隐私风险也在增加，因为 HMAS 元数据可能包含敏感的私人信息。我证明，基于个体微生物组的分类群特征以及 HMAS 数据的汇总统计信息的简单检验统计量可以揭示个体微生物组在 HMAS 样本中的成员身份。特别是，从小规模 HMAS 获得的物种级分类数据可能非常容易受到隐私风险的影响。建议为 HMAS 数据隐私制定最低准则，并建议在研究设计时使用所提出的模拟方法评估 HMAS 隐私风险。
