Intersection of DNA privacy and whole-genome sequencing.

In 2012, the US Presidential Commission for the Study of Bioethical Issues issued a report on Privacy and Progress in Whole Genome Sequencing (1 ). This report made recommendations on the balance between research advancements and privacy risks derived from whole-genome sequencing. Some of the recommendations from the Commission included strong data protection, informed consent, and the establishment of consistent genomic privacy protection between states. The Commission’s recommendation for consent before whole-genome sequencing came with exceptions for legitimate law enforcement purposes. Indeed, the report specifically recommended that “[o]nly in exceptional circumstances should entities such as law enforcement or defense and security have access to biospecimens or whole genome sequence data for non health-related purposes without consent” (1 ). This authorized use of genomic data by law enforcement should be considered in the context of the 2013 revelation of the US National Security Agency’s secret data collection and surveillance program of Internet users (2 ). In the near future, publicly available whole-genome sequencing data may be similarly examined and surveyed by law enforcement or government agencies. On one end of the DNA privacy spectrum are the millions of DNA short tandem repeat (STR) identification profiles in government, health care, and commercial databases. These STR identification profiles are characterizations of small regions of DNA sequence data that contain sufficient information to identify individuals or relatives of individuals by matching a set of maternal and paternal alleles. This powerful identification tool has been used for several decades in the setting of bone marrow transplantation, law enforcement, paternity lawsuits, and military identification. On the other end of the DNA privacy spectrum is the emergence of whole-genome sequencing; these data sets contain both protein coding and noncoding DNA, including STR regions. Genomic data sets contain information on current and future diseases and are essential for the future discovery of genetic etiologies of disease. The Commission noted that although a person’s genomic data is unique, the data could be deidentified by removing the link between the data and common identifiers (e.g., name, date of birth, home address). In this scenario, reidentification of the person would be difficult without a method of linking back to the common identifiers. The analogy used by the Commission was that fingerprints are unique, but they are useful only when the fingerprint information exists in a database linked to the common identifiers of a person. The Commission stated that although whole-genome sequence data are “uniquely identifiable,” they are not currently “readily identifiable” (1 ). Thus, genomic data is unique for each person, but does not automatically identify an individual. In their 2012 report, the Commission did not extend the fingerprint analogy to recognize the possibility that existing databases of DNA STR identification profiles may be used as a link between genomic data and a person’s common identifiers. Sequencing technologies and software algorithms are emerging that make it possible to derive STR profiles from whole-genome sequencing data sets. Publicly available whole-genome data sets may become a resource for the reidentification of individuals or their relatives by linking their whole genome– derived STR profiles with common identifiers in DNA STR identification profile databases. In the US, DNA STR profiling has become commonplace in forensics and law enforcement. Thirteen specific STR loci are used by the US Federal Bureau of Investigation (FBI) Combined DNA Index System (CODIS). Currently, 11 million individual CODIS STR profiles are held in tightly controlled databases with access limited to authorized individuals. In the case of CODIS, STR profiles are not only from convicted offenders, arrestees, and missing persons, but also from the departments of Defense and Homeland Security. The CODIS database contains only the STR profiles and not the common identifiers of a person; however, submitting entities maintain the data linking common identifiers to STR profiles. Separate from the CODIS database are also 1 Nationwide Children’s Hospital, Columbus, OH; 2 Department of Pathology, Children’s Medical Center and University of Texas Southwestern Medical Center, Dallas, TX; 3 Eugene McDermott Center for Human Growth and Development and 4 Quantitative Biomedical Research Center, University of Texas Southwestern Medical Center, Dallas, TX. * Address correspondence to this author at: Department of Pathology, Children’s Medical Center andUT SouthwesternMedical Center, 1935Medical District Dr, Dallas, TX 75235. Fax 214-456-4713; e-mail [email protected]. Received November 21, 2014; accepted December 18, 2014. Previously published online at DOI: 10.1373/clinchem.2014.235499 © 2014 American Association for Clinical Chemistry 5 Nonstandard abbreviations: STR, short tandem repeat; FBI, US Federal Bureau of Investigation; CODIS, Combined DNA Index System; ICGC, International Cancer Genome Consortium. Clinical Chemistry 61:7 900–902 (2015) Perspective