Identity inference of genomic data using long-range familial searches

Area of Science:

• Genomic privacy
• Forensic genealogy
• Bioinformatics

Background:

• Consumer genomics databases now contain millions of individuals' genetic information.
• Law enforcement agencies are increasingly using these databases for suspect identification through familial matching.

Purpose of the Study:

• To investigate the effectiveness and reach of using consumer genomics databases for identifying individuals.
• To assess the potential for implicating a significant portion of the population, particularly those of European descent.

Main Methods:

• Analysis of genomic data from 1.28 million individuals tested via consumer genomics.
• Projection of identification success rates based on familial matching thresholds (e.g., third-cousin or closer).

Main Results:

• Approximately 60% of searches for individuals of European descent are projected to yield a third-cousin or closer match.
• The technique has the potential to identify nearly any U.S. individual of European descent in the near future.
• Research participants in public sequencing projects can also be identified using this method.

Conclusions:

• The widespread use of consumer genomics poses significant privacy risks, enabling identification through familial DNA.
• A mitigation strategy and policy implications for human subject research are proposed in response to these findings.