The legal and ethical protection of personal data depends largely on the potential to identify the individuals behind the data. Following a widespread assumption, genetic data are usually considered inherently self-identifying. However, this view likely simplifies the varying degrees in identifiability and ignores specific context. In practice, the risk of re-identification largely depends on the type and amount of data and accompanying metadata that is collected.  

As part of the ongoing project "Forensic Genetics Frequency Databases: Between Science and Law Enforcement", funded by the Swiss National Science Foundation, we investigated the identifiability of Y-STR haplotypes and mitochondrial sequence variants registered on YHRD and EMPOP through an empirical re-identification attempt. The purpose was to assess the actual risk of re-identification associated with these publically available data types.  

Our findings challenge the assumption that all genetic data are inherently identifying and demonstrate the need for more nuanced definitions of anonymity and re-identification risks in genetics. In addition, the question arises of when a narrowing down to a specific population can be considered a group identification and how elaborate and practically applicable the concept of group identification actually is. Proper definitions and classifications are crucial for the evaluation of research projects, ethical oversight and the development of appropriate data protection policies. The results from this project can have implications for future data governance in Forensic Genetics. 

Authors

• Alexandre Gouy (AlgoLife SARL, France) 
• Martin Zieger (University of Bern, Switzerland)