Forensic Fakery - Fabricated DNA Evidence and Countermeasures
It is surely no exaggeration to suggest that DNA analysis has revolutionized the pursuit of justice. Indeed, according to a 2009 report by the National Academy of Sciences, DNA analysis is the only method commonly used by forensic labs that has real scientific verification of its power to match suspects with crimes. It is this power that has led to a popular impression of the infallibility of forensic science in general and DNA evidence in particular, both in terms of its ability to exonerate the innocent and convict the guilty (Thompson 1997.) This perception has surely been reinforced by popular criminal-forensics TV dramas such as "CSI" (and its two associated spinoffs), "NCIS", "Bones", "Crossing Jordan", "Cold Case", and other shows that take the form of the hoary "police procedural" genre updated with modern science and technology.
In July 2009, researchers at Israeli life science company Nucleix announced in a paper a method for the production of artificial DNA samples that would be indistinguishable from legitimate samples under standard forensic protocols for collection and analysis of DNA evidence (Frumkin et al, 2009), a method that turned the FBI's CODIS database - a national database of DNA profiles developed from Short Tandem Repeat (STR) analysis - against itself. With a relatively small "toolbox" of samples, researchers were able to construct a sample of DNA that would match a specific CODIS profile when analyzed for STRs. With PCR cloning technology widely available and cheap, a dastardly geneticist could easily synthesize an arbitrary amount of DNA and apply it to the crime scene, perhaps to cover up his own involvement or even implicate an unrelated party. In this paper I will attempt to summarize the Nucleix procedure and countermeasures suggested by the researchers and by others, and discuss the implications of this troubling technology on criminal justice matters.
The Nucleix Procedure - "Biological Identity Theft"
Researchers Dan Frumkin, Adam Wassertstrom, Ariane Davidson of Nucleix, along with Arnon Grafit of the Israel police, demonstrated a method to fool forensic investigators with false DNA evidence that exploits the standard forensic reliance on Short Tandem Repeat (STR) profiling and the FBI's Combined DNA Index System (CODIS), a national computerized database of DNA profiles assembled by Federal, state, and local law enforcement agencies. CODIS assembles profiles according to polymorphisms at 13 STR sites. A conspirator armed with a sufficiently wide variety of DNA samples would have in his possession a sample for every known polymorphism of each of 13 STR sites. Since traditional STR analysis with CODIS is merely a matter of determining the specific polymorphisms at these 13 loci, the conspirator could "mix and match" a selection of selective amplicons that, when mixed together, would give the appearance of matching a specific STR profile. If our conspirator could obtain a specific individual's STR profile - say, by access to CODIS through surreptitious means, or even directly by the same sort of surreptitious DNA collection methods used by police investigators, like collection of cigarette butts, drinking cups, or other detritus - he would have all the information he needed to synthesize an arbitrary amount of genetic material that, when analyzed by standard forensic DNA analysis protocols, would positively match that target individual's actual STR profile.
While this may sound like a substantial amount of difficult, highly-trained work, that may not be the case. A four-year bachelor's degree in any of the life sciences would doubtless expose someone to all the necessary training, equipment, and concepts required to sample DNA, perform polymerase chain reaction (PCR) amplification, and combine samples to construct an artificial profile. Additionally, the equipment and materials to sample DNA, store samples, and perform PCR amplification are neither expensive nor uncommon; even a local community college is likely to have access to a relatively inexpensive PCR thermocycler. The Nucleix researchers estimated that it would take no more than 425 different DNA samples to construct any desired STR profile. And it exposes a startling weakness in the CODIS doctrine - while a massive national database of DNA profiles has proven an incredibly useful resource for closing cases and identifying suspects, the broadness of the database - many states are even collecting DNA evidence from arrestees - means that it contains millions of potential "patsies" for the determined conspirator.
In a startling confirmation of the efficacy of Nucleix procedure, researchers collected samples from test subjects and constructed artificial samples using the procedures described, then shipped the samples to a prominent forensics lab for testing. The lab was unable to discern the difference, matching individuals to the artificial samples constructed to match their profiles from DNA from completely different sources. Further, analysis with a popular profiling system called Profiler Plus detected no anomaly and returned the same fictitious matches. The implication is clear - the standard suite of protocols and techniques currently employed by forensics investigators is insufficient to detect DNA falsification of this type.
The Nucleix Countermeasure
Fortunately the same investigators that exposed this weakness have suggested a countermeasure, in the form of detection of DNA methylation. In vivo, nuclear DNA becomes methylated at cytosine bases by the addition of a methyl group to the pyramidine ring (Nelson and Cox 2004.) This is a naturally occurring process that, in the living organism, is involved in gene expression and regulation, as well as DNA replication (Nelson and Cox 2004.) DNA amplified by PCR isn't subject to this sort of regulation, and as a result is not methylated. Lab assays to detect methylation are available, but as yet, the procedure is not well-automated, is time-consuming and laborious, and is not frequently included as part of a forensic analyst's training (Cottrell 2004.) Further, since "faked" DNA evidence gives every appearance of being legitimate (aside from its lack of methylation), and it may not be obvious in which cases a life sciences graduate may be involved, methylation assay must be performed on every forensic DNA sample if we are to retain our confidence - legal and moral - in DNA profiling as a criminal justice tool. Fortunately the wide publicity of the Nucleix article is having an effect. Several life sciences companies have announced development of more rapid and automated assays, Nucleix among them (Cottrell 2004, Eads 2000).
Anything that shakes our confidence in DNA evidence also shakes our confidence in the convictions (and exonerations) resulting from that evidence. It is likely that the Constitution's prohibition against "double jeopardy" will protect the exonerated from suspicion cast on the DNA evidence that cleared them. And while the technology and training to fake DNA evidence is hardly rare, faked DNA evidence has never been detected in any criminal case. Nonetheless it is not difficult to imagine a flood of appeals from those convicted mostly on the strength of DNA evidence asserting that they were "framed." Indeed it has not been unknown for forensic investigators to falsify reports or DNA analyses to frame individuals with the assistance of the police (Kelly 1998).
But the flip side of a loss of public confidence in the capacity of law enforcement personnel is a greater scrutiny towards the collection, management, and interpretation of that evidence. If juries can be relied upon to skeptically scrutinize the evidence and testimony of forensic experts, instead of taking such testimony as holy writ, abuses like the FBI Crime Lab scandal (Kelly 1998) can be avoided. Ironically the loss in confidence in the "infallibility" of forensic DNA evidence may give us greater confidence in criminal trial proceedings, knowing that convictions or exonerations are the result of a reasoned assessment of forensic evidence, along with its strengths and weaknesses, not exaggerations of forensic power.
Frumkin, D., Wasserstrom, A., Davidson, A., Grafit, A. Authentication of Forensic DNA Samples. Forensic Science International: Genetics - 17 July 2009 (10.1016/j.fsigen.2009.06.009)
Nelson, D., and Cox, M., 2004. Lehninger Principles of Biochemistry, Fourth Edition. W. H. Freeman Press, New York, NY.
Thompson, William C. Sociological Perspective on the Science of Forensic DNA Testing. A; 30 U.C. Davis L. Rev. 1113 (1996-1997)
Cottrell SE, Distler J, Goodman NS, Mooney SH, Kluth A, Olek A, Schwope I, Tetzner R, Ziebarth H, Berlin K. A real-time PCR assay for DNA-methylation using methylation-specific blockers. Nucleic Acids Res. 2004 Jan 13;32(1):e10.
Eads CA, Danenberg KD, Kawakami K, Saltz LB, Blake C, Shibata D, Danenberg PV, Laird PW. MethyLight: a high-throughput assay to measure DNA methylation. Nucleic Acids Res. 2000 Apr 15;28(8):E32.
Kelly, J. and Wearne, P. Tainting Evidence: Behind the Scandals at the FBI Crime Lab. Simon and Schuster, New York, NY.