Those who transition from haplotype ancestry maps of geographically isolated populations to techniques for drawing human physical properties from a DNA sample are making many scientifically contested assumptions. They use a process called Ancestry Informative Marker (AIM) technology. According to Fullwiley,
AIMs-based technologies, like DNA Witness, are attempts to model human history from a specifically American perspective to infer present-day humans’ continental origins. Such inferences are based on the extent to which any subject or sample shares a panel of alleles (or variants of alleles) that code for genomic function, such as malaria resistance, UV [ultraviolet] protection, lactose digestion, skin pigmentation, etc. There is a range of such traits that are conserved in, and shared between, different peoples and populations around the globe for evolutionary, adaptive, migratory, and cultural reasons. To assume that people who share, or rather co-possess, these traits can necessarily be “diagnosed” with a specific source ancestry is misleading.61
Without a cold hit in CODIS the police must resort to traditional policing techniques, which certainly would not include rounding up all redheaded persons with green eyes. But if familial searches are used in CODIS to troll for suspects (see chapter 4), then these gene-to-phenotype correlations may, in the minds of criminal investigators, play a role in narrowing the search. Suppose that a familial search yielded 20 current or past prisoners or, in some states, arrestees. If all the partially matched profiles had additional genetic information accessible to police, then the police could troll through the 20 saved biological samples for a person with red hair and green eyes who was a partial match in the familial search. If the technique described by Frudakis shows increasing promise, there will be pressure to maintain biological samples and to transform CODIS from a DNA profile identity database to a data bank for building physical profiles of individuals from their DNA. Fullwiley believes that these techniques are substantiated neither in ancestry testing nor in forensic phenotyping of DNA: “As a forensics market version of the AIMs technology, DNA Witness may offer precise mathematical ancestry percentages, but the accuracy of that precision remains debatable.”62
Because the national forensic DNA databases have a disproportionately high percentage of African Americans, the traits most common to this group will be sought out in developing profiles of suspects. Moreover, there will be pressure to use the sample population of African Americans in CODIS to refine the genotype-to-phenotype correlations. Although the developers of “molecular photofitting” agree that it is not foolproof, they believe that “there is information about M [pigmentation] we gain by knowing an individual’s genomic ancestry and as long as it is communicated responsibly, this information would be more useful for a forensic investigator as a human-assessed measure of skin color would be.”63 In other words, Frudakis claims that this method is more reliable than eyewitness testimony of skin color while also providing data with a very poor correlation between pigmentation and ancestry.64
The forensic DNA data banks in the United States and the United Kingdom contain a disproportionately high number of people of color because of the demographics of crime and criminal arrests. Even if the purpose of DNA data banks remains true to its original mission of matching the identity of samples, the significant racial skewing of the database, combined with a well-established systemic racial bias in the overall system, means that people of color who commit crimes are more likely to be identified in a DNA match, while white individuals will be more likely to escape detection. The expansion of databases to arrestees will worsen this situation, since decisions concerning who is arrested are highly discretionary and therefore especially prone to bias. Since minorities are arrested disproportionately to their contribution to criminal activity, these communities will be placed under greater DNA surveillance and subjected to greater stigmatization.
In an effort to develop “profiles” of perpetrators from DNA left at a crime scene, the criminal justice community will inadvertently reify race as a scientific term, even though it has been widely discredited and undefined scientifically. From this, new consequences of abuse are likely to result. One of these consequences involves picking up the usual minority suspect from weak probabilistic assignments of skin color from DNA through ancestry analysis. Skin pigmentation is a continuous spectrum from albino to dark brown and every shade in between, with no clear breaks. Race, however, is socially constructed as a two-value concept—black or nonblack (white). When DNA is used to determine pigmentation, the outcome, as interpreted by police investigators, will most likely be that the perpetrator was black or nonblack. If the science remains weak, this type of inference can exacerbate the racialization of criminal justice because society still operates under the one-drop rule. If the forensic investigator estimates that a DNA sample from an individual shows 90 percent European and 10 percent African ancestry, does the one-drop rule imply that he or she should be looking for a person of color? Or does the melanin index (with its poor predictive power) trump ancestry? Although the introduction of these techniques in the courtroom is unlikely, there are no restrictions on their use by police in generating suspects. That is where we can see the confluence of expanded data banks of innocent people, disproportionate numbers of minorities in the data banks, and the use of not-ready-for-prime-time science to invade the privacy of people’s lives.
Chapter 16
Fallibility in DNA Identification
It will not be easy to dispute the prevailing wisdom, fed by CSI-style media fantasies, that forensic science is virtually infallible. Yet, the intellectual weaknesses of many of the “forensic sciences” are now becoming increasingly apparent.
—William Tobin and William Thompson1
It can justifiably be argued that, to date, the greatest contribution to civilization arising directly from the genetics revolution has been the exoneration of hundreds of falsely convicted individuals who have logged thousands of years in prison, some on death row. The failure to match the DNA of the alleged perpetrator with the DNA at the crime scene has, in most cases, trumped other physical and circumstantial evidence, including eyewitness testimony, that juries deemed “beyond a reasonable doubt” in deciding the guilt of an innocent individual. For many, DNA exculpatory evidence has come to mean “beyond a conceivable doubt.”
Although DNA exonerations are hailed as one of the most important achievements of forensic science, the primary goal of the government in applying DNA to criminal justice has been one of finding guilt, not innocence. District attorneys have deployed DNA as indisputable evidence to argue that a suspect was the perpetrator of a rape or was at the scene of another type of violent crime. DNA evidence has sent thousands of individuals to prison, but it has only helped exonerate around 250 falsely convicted individuals.2 However, in the process of vetting suspects and evaluating alibis, police also use DNA to exclude individuals in routine casework before they are drawn too deeply into the web of investigation and surveillance. Early exclusions are a win-win situation for criminal justice and those individuals who might otherwise have to experience protracted periods under the cloak of suspicion and with fear of prosecution.
When DNA evidence is used to support a hypothesis either of guilt or of innocence, and all other hypotheses that contradict the proffered one are either patently false, are highly improbable, or lack any substantial evidentiary support, DNA evidence takes on authoritative preeminence. It is under these conditions that DNA evidence is presented in the popular Crime Scene Investigation (CSI) vernacular as infallible. But in reality, ideal circumstances are rarely the way in which criminal cases get played out. DNA infallibility is a myth, even though DNA evidence can be highly authoritative and effective in identifying and prosecuting criminals. The following describes eight central myths of forensic DNA evidence. These myths have been identified from media accounts and in exaggerated claims made by individuals who view DNA as the ultimate and incontestable authority within forensic science.
Myth of DN
A Consistency
DNA Is DNA Is DNA. All DNA Evidence Is Strong Evidence; There Is Not Much Difference in Quality from One Sample to the Next.
Although DNA is revered as the “gold standard” of forensic science, not all DNA is the same. William C. Thompson has pointed out that there is considerable case-to-case variation in the nature and quality of DNA evidence.3 For example, it is not uncommon that biological evidence collected at the scene of a crime does not provide a full DNA profile when analyzed. Sometimes only a few cells are picked off a piece of clothing or an object. Also, in many cases DNA is analyzed or reanalyzed many years after a crime has been committed, and by that time it has further degraded. DNA, like any other chemical, will break down over time. The rate of degradation depends on the type of cells (saliva, blood, semen, skin), as well as the conditions under which the samples are stored. Improper storage or exposure to sunlight, moisture, bacteria, or other unfavorable conditions can accelerate degradation rates. Where significant degradation has occurred, the DNA analysis may not result in a full DNA profile.
Some DNA evidence is also less probative than other DNA evidence. For example, a DNA sample taken from a cigarette butt on the street where a crime was committed is less likely to have come from the perpetrator than DNA extracted from a vaginal swab of a rape victim.
DNA evidence presented in a case is only as good as the DNA found at the crime scene. When DNA evidence is compromised because the biological sample produces less than a full profile or because it may be unrelated to the crime in question, that should be acknowledged from the start. Partial DNA matches that occur as a result of degradation must be understood to carry far less weight than full matches where DNA is being used to establish that an individual was at the scene of a crime. Likewise, matches with DNA that might have been left by an innocent passerby should not carry the same relevancy as those where the DNA evidence is more likely to have come from the perpetrator.
Myth of Infallible Matches
There Are No False Positives in DNA Testing. If Two Samples of DNA Are Found to Match, Then the Samples Must Have Come from the Same Individual.
Although current scientific consensus supports the conclusion that, except possibly for identical twins, no two individuals have an identical genome, the conclusion of individuated genomes does not imply that a match of two DNA profiles indicates that the samples came from the same individual. False positives can and do occur in forensic DNA analysis. They can happen because of error, contamination, interpretation of the output of DNA analyzers, and chance profile matches that can be expected in a sufficiently large population.
We have explained in some detail in chapter 1 the stages involved in performing DNA analysis for identification. The reliability of the process depends on the quality of the DNA obtained at the crime scene; the care with which it is collected, labeled, and transported; the standards and quality-control procedures of laboratories performing the DNA profile analysis; and the interpretation of the DNA analyzer data, including whether a partial profile or a mixed profile is obtained. There are a number of opportunities where errors can occur in the collection, handling, and storage of DNA samples that can result in false positives and, therefore, constitute a risk of incriminating an innocent person.
Sample mix-ups and mislabeling in rape cases, where biological evidence that is being compared contains mixtures of the perpetrator’s DNA with that of the victim, can lead to the incrimination of the wrong person. If the reference DNA samples are switched, then the DNA that is believed to have come from the suspect (but in actuality is that of the victim) will invariably be found to be included in a vaginal swab. In 2000 the Philadelphia City Crime Laboratory admitted to having accidentally switched the reference samples of the defendant and the victim in a rape case. As a result of the sample switch, the lab issued a report that stated that the defendant’s profile was included in a mixed sample taken from vaginal swabs. The report also stated that the defendant was a potential contributor of what the analysts took to be “seminal stains” on the victim’s clothing, which they later realized were in fact bloodstains from the victim.4
BOX 16.1 The Case of Lazaro Soto Lusson
In 2002 it was discovered that 26-year-old Lazaro Soto Lusson was mistakenly charged with multiple felonies because the Las Vegas Police crime lab switched the labels on two DNA samples. While in jail on an immigration hold, Lusson’s cellmate, Joseph Coppola, accused him of rape. Police took DNA samples from both men to investigate the allegation. While they were conducting the analysis, they ran the samples against the state database and matched Lusson’s mislabeled DNA to two unsolved sexual assaults. Lusson faced life in prison and was incarcerated for over a year before this mistake was discovered.
Source: Glenn Puit, “Wheels of Justice Turn Slowly,” Las Vegas Review-Journal, July 6, 2002.
DNA samples can also be contaminated, either before or after collection, especially if they are not stored under proper conditions. Samples can be contaminated by the inadvertent transfer of trace amounts of DNA. Ironically, this error type is of increasing concern as DNA tests become more sensitive.5 Lab analysts have cross-contaminated samples by not properly sterilizing lab equipment between cases or with their own DNA by not wearing the proper protective clothing while conducting DNA analyses.
Even trace amounts of outside DNA can complicate a DNA analysis. Where more than one source of DNA is present in a mixture, the results of the DNA analysis can be difficult to interpret. It can be difficult to tell how many individuals contributed to the source of the DNA sample, let alone which alleles are associated with each of those contributors.6 The presence of one source of DNA can also mask another, and degradation might cause one source to go undetected. Any of these situations can lead to a false positive match.
BOX 16.2 The Case of Timothy Durham
In 1993 Timothy Durham was convicted of raping an 11-year-old girl and sentenced to 3,000 years in prison despite having produced 11 alibi witnesses who placed him in another state at the time of the crime. The prosecution’s case rested almost entirely on a DNA test, which showed that Durham’s genotype matched that of the semen donor. Postconviction DNA testing showed that Durham should have been excluded as a possible suspect, and reanalysis of the initial test showed that the misinterpretation arose from the difficulty of separating mixed samples. The lab had failed to separate completely the male and female DNA from the semen stain, and the combination of alleles from the two sources produced a genotype that could have included Durham’s. Durham was released in 1997 after serving four years in prison.
Source: W. C. Thompson, F. Taroni, and C. G. G. Aitken, “How the Probability of a False Positive Affects the Value of DNA Evidence,” Journal of Forensic Sciences 48, no. 1 (January 2003): 47–54, information at 50.
Sample cross-contamination appears to be a surprisingly common lab error. Thompson has found that these errors are chronic and occur even at the best-run DNA labs.7 Under a guideline issued by the FBI, DNA laboratories are required to maintain corrective-action files to keep track of discrepancies that arise in casework. Many laboratories do not adhere to this guideline, but Thompson has reviewed corrective-action files for some labs where a file is maintained. A small laboratory in Bakersfield, California, for example, documented
multiple instances in which blank control samples were positive for DNA, an instance in which a mother’s reference sample was contaminated with DNA from her child, several instances in which samples were accidentally switched or mislabeled, an instance in which an analyst’s DNA contaminated samples, an instance in which DNA extracted from two different samples was accidentally combined into the same tube, falsely creating a mixed sample, and an instance in which a suspect tested twice did not match himself.8
Thompson worries that this and other similar examples are only “the tip of the iceberg,” especially since these represent only the errors that the lab itself has caught, corrected, and documented.
In 2004 the Seattl
e Post-Intelligencer reported that forensic scientists at the Washington State Patrol Laboratory had made mistakes while handling evidence in at least 23 major criminal cases over three years. Most of these mistakes involved contamination by DNA from unrelated cases, from the lab analysts themselves, or between evidence in the same case.9
Cross-contamination of DNA samples in laboratories has led to false cold hits in several cases. For example, in Washington State a cold hit turned up when DNA from a rape case was compared with the state database. However, the juvenile offender who appeared to be the source of the DNA would have been only 4 years old at the time the rape was committed. In realizing that this individual could not have been connected with the crime, the Washington State Patrol Laboratory came to understand that the juvenile’s sample had been used as a training sample by another analyst when the rape case was being analyzed.10
Processing DNA samples requires that humans collect and handle biological samples, which are then subjected to laboratory techniques run by human technicians. DNA testing is only as reliable as are the people overseeing each of these processes, and infallibility simply cannot be achieved. Therefore, forensic scientists must depend on quality control, retesting, troubleshooting, and transparency of every decision made in the process to achieve reliable, trustworthy forensic evidence every time.
Genetic Justice Page 34
