Please, Let’s Bury the Junk: The CODIS Loci and the Revelation of Private Information
By D.H. Kaye[*]
In a recent essay, Professor Simon Cole asks “Is the ‘Junk DNA’ designation bunk?”[1] He concludes that in one sense, it is not. There is no scientific evidence that the specific DNA variations used to identify the sources of crime-scene DNA perform any biological functions. Nonetheless, he contends that this fact, in and of itself, does not obviate the concern that the specific STR profiles stored in law enforcement databases of offenders (and sometimes arrestees) might be used to extract medically or socially sensitive information. I agree and have said as much in the past.[2]
Professor Cole also writes that “[t]he
privacy threat posed by forensic STRs may not be great,”[3] but he does not explain the basis for this
view, and many of his remarks could be construed as being more consistent with
the opposite conclusion—that the privacy threat may well be great. He criticizes the assurances of forensic
scientists and human geneticists that, at present, “forensic DNA has no predictive
value or medical significance”[4]
as “misleading” and “not fully informative.”[5] He proposes that the records of the STR types of offenders contained in existing
law enforcement databases “may, in fact, be precisely the kind of ‘predictive
medical information’ that concerns privacy advocates,”[6] and he refers to STRs as potential “markers”
having “predictive utility.”[7] In
particular, he asserts that “the forensic STRs . . . correlate with . . .
disease-causing genes”[8] and “phenotypically perceived race.”[9] He
concludes that “[i]f some forensic STRs are correlated with genes that cause
physical traits, . . . the public can [and should] be informed of that fact”[10] so that it “can decide for itself whether
and to what extent the privacy risk offsets the benefits of genetic databases.”[11] The genetically influenced physical traits
that he proposes are discernible from the DNA sequences used in criminal
identification databases in the United States include diseases that would be of
interest to insurance companies or employers and physical features associated
with conventional racial categories.
These remarks require clarification. Just as the argument that nonfunctional DNA
cannot be a threat to privacy is superficial, it would be incomplete and
misleading simply to inform the public that an STR profile contains information
that is correlated to physical traits such as disease and possibly behavioral
predispositions and hence could be used to predict whether an individual will
develop a disease. By innuendo, this
formulation suggests that these nonfunctional loci, which are very weakly associated
(if at all) with disease or behavior, are comparable to the loci used in much
more powerful modern genetic testing for the DNA sequences of mutations that do
cause disease.
This Colloquy Essay therefore analyzes in
greater depth the medical and biological implications of the DNA records in the
National DNA Index System (NDIS) and its local and state components. It explains why the STR profiles are useless
as a “genetic test to screen for any particular disease.”[12] No
one can say for certain what the future of genetics holds, but based on current
knowledge and practice, the information coded in the databases is and will
remain, with the limited exceptions noted below,[13] useful only for identification.
To develop these points, Part I briefly
describes the four possible ways in which genetic loci could possess predictive
or diagnostic value with regard to diseases and explains why these mechanisms
have not led, and probably cannot lead, to useful screening tests with the
Convicted Offender DNA Index System (CODIS) profiles in national, state, and
local databases. Part II considers the
“physical traits” and familial relationships that the CODIS STRs can be used to
identify. That the profiles carry
limited information about an individual’s race and familial relationships has
long been part of the public dialogue, and Part II places the resulting privacy
issues in perspective. Part III comments
on analogies between STR types and fingerprints, social-security numbers, and
the like, employed when discussing these issues in the public forum.

