Forensic DNA Profiles: Database Expansion, Familial Search, and a Radical Solution

This paper outlines the substantial privacy risks posed by the human genome, and explains how sensitive information is omitted from FBI DNA profiles. It then details the quiet expansion of law enforcement DNA databases, including the controversial technique of “familial DNA search,” and argues for a population-wide, universal DNA database. Genomes, Profiles and Privacy Risks The human genome is potentially a privacy nightmare. Citizens and scholars alike are increasingly concerned about the threat of highly personal information being exploited to invade individual privacy. Genetic privacy is worthy of concern: the genome is an information-rich data warehouse that poses a considerable privacy threat for two reasons. First, human genomic DNA encodes the biological blueprint necessary to create an individual person. As such it necessarily includes a vast amount of sensitive information, from outwardly observable characteristics, to sensitive health information such as genetic predispositions to disease. Indeed, it is difficult to imagine any information more deeply personal. Moreover, the genome carries substantial information about relatives, particularly close relatives like parents, children and siblings. Second, we all routinely shed copies of our genomic DNA. A human being consists of some 50-100 trillion cells, almost all of which carry a complete copy of that individual’s genome. And we deposit copies of our genomes everywhere—in shed skin cells, in saliva on soda cans and cigarette butts, and so on—free for the taking (and analysis). This reality will surely pose challenging legal questions in coming years. However, this paper concerns a particular and very narrow use of genomic information: the forensic DNA profile, as used in Combined DNA Index System (CODIS), Copyright © 2009, Michael Seringhaus. Used with permission by Association for the Advancement of Artificial Intelligence (www.aaai.org). All rights reserved. the FBI DNA database system. These profiles ignore all the sensitive genomic information mentioned above, and distill just 26 two-digit numbers from any genomic sample. As such, they pose little if any privacy risk, and accordingly different standards should be brought to bear in evaluating their collection and retention by government. This paper presents the central argument of a longer paper currently in preparation: namely, the argument for a population-wide, universal database of CODIS-style DNA profiles. The Genome: Scientific Background Your genome is composed of some 3 billion nucleotides (chemicals represented by the letters A, C, G and T) divided among 23 chromosomes. Like all humans, you are a diploid organism—meaning that each somatic cell in your body carries two copies of each chromosome—and thus you have 46 chromosomes per cell. In sequence terms, your genome is roughly 99.5% identical to that of any other human. That is, when our genome sequences are aligned, you and I should differ at only about 0.5% of our 3 billion sequence positions, or 15 million nucleotides in all. It appears that only a small fraction of the genome, about 2%, encodes genes. (In general terms, genes are nucleotide sequences that are “read” to create cellular components called proteins.) The remaining 98% consists of what was once called “junk DNA.” (No more: this term has fallen out of favor in recent years as researchers uncovered functionality in this vast repository of noncoding DNA.) Nonetheless, it remains true that the bulk of the genome does not directly encode genes, and whatever other functions this sequence may perform, it almost certainly also includes vast stretches of true “junk”: meaningless repeated DNA left over from thousands of years of viral invasion and duplication. Despite being freighted with seemingly meaningless extra sequence, the genome is your biological blueprint, and an immensely potent information repository. When decoded by cellular machinery, the genome provides