Read any good books lately?

Let the DNA Fit the Crime by Richard C. Lewontin

Hercule Poirot has an easy time of it. He needs only the logic provided by his “little gray cells” to discern, unerringly, which of the dozen or so guests at the country house or passengers in the sleeping coach have committed the crime, nor is there any doubt that those responsible would be found guilty were a trial to be held. But what is a real-life New York investigative police officer to do when, in a city with a population of four million males, a savage rape has left its victim lying unconscious in a remote spot, permanently unable to recall any detail of the attack? One obvious direction of investigation is to connect it with some other violence that took place near the same place at around the same time.

At around 9:00 in the evening of April 19, 1989, a young woman entered Central Park near the Metropolitan Museum and headed north intent on an evening jog. At about the same time a large group of Latino and black teenagers intent on a macho adventure entered the park at its northeast corner and headed south. It seems very likely from the layout of the paths in the park that the boys’ southward route overlapped with the jogger’s northward way.

After a number of aggressive encounters with passersby including one serious beating, the group slowly broke up and five of them left the park on the west side where they were arrested by the police, to whom their behavior had been reported. Much later, at about 1:30 AM, the blood-drenched body of the raped, half-conscious Patricia Meili was discovered by two passersby not far from where two of the gang’s aggressive encounters had been reported. She has never been able to recall the details of the attack or her attackers. The police investigators made what must have seemed to them an obvious connection. The “Central Park Five” became the center of the rape investigation.

In dealing with the suspects, the police used the classic “good cop, bad cop” technique. The boys were alternately shouted at and treated sympathetically in order to get them to confess. They were promised that if they confessed they would be allowed to go home. Eventually they all confessed to having observed the rape, either from afar or having held the woman down while another member of the group actually raped her, but none admitted to the rape itself. Their stories were inconsistent with one another and with the actual condition in which the victim’s injured body was found. For example, despite her bloody condition, there were no bloodstains on any of the accused’s clothing. It seems clear that these immature boys, desperate to escape the conditions in which they were held, and perhaps reasoning that they would not, in the end, be held responsible for something they did not do, were willing to give the police what they wanted.

In this they were mistaken. In preparing for their indictment, the prosecutors arranged for a test of the DNA samples taken from Meili’s cervical swab. Only a weak pattern was found, but it did not match the DNA of any of the five defendants. On learning of the result the lead prosecutor told one of the other attorneys that “I feel I have been kicked in the stomach,” but that did not cause her to look elsewhere for the rapist. Then, in the middle of the trial, the New York Police Department’s serologist, Mary Veit, found a much larger semen stain on one of Meili’s socks and when the DNA was tested this showed conclusively that none of the defendants had left it.

Nevertheless, the prosecution persisted and, based on the confessions, convinced the jury to convict all five. They spent between six and thirteen years in prison. While the last to be released was still serving his term, he was approached by another prisoner, Matias Reyes, who confessed to being the real Central Park attacker, and whose DNA, when tested, showed him to be a serial rapist, responsible for a large number of unsolved attacks. Nevertheless, various members of the New York Police Department and the district attorney’s office continued to insist that justice had been done, that Reyes’s confession was false, and that the DNA evidence was inconclusive. Still, the convictions were vacated in 2002 and the young men have sued the city of New York for “malicious prosecution, racial discrimination, and emotional distress.” The city is contesting their lawsuit.

This case is not unique, even in its details. In 1991 in Chicago, five young men, the “Dixmoor Five,” were accused of the rape and murder of a fourteen-year-old girl and were convicted after their confessions to the police, despite exculpatory DNA evidence. Only in 2011 did the state’s attorney move to vacate the convictions on the basis of the DNA test. Three years later a prostitute was raped and strangled on the South Side of Chicago and four boys were convicted of the crime after confessing to police interrogators who promised to let them go home if they confessed. One of them describes a familiar situation:

I was 17, I’d never been in any type of trouble like that…. I didn’t know the weight or the magnitude of what a confession could do. It cost me 17 years of my life.

Only last summer, a long-belated examination of DNA evidence pointed, in fact, to a known violent sexual offender who was actually seen in the neighborhood where the body was discovered, and interviewed by the police at the time. Nevertheless, the same state’s attorney had, as of November 15, no plans to dismiss the convictions because “DNA evidence in and of itself is not always the ‘silver bullet’ that it is sometimes perceived to be.” It may not be irrelevant that the probable rapist is now dead so that his trial and conviction are impossible. The exculpation of the convicted youths would then leave the case officially and unsatisfactorily “unsolved” and the state’s attorney with one less success to support her record in office.

The increasing importance that evidence from DNA samples now has in the criminal justice system makes it essential that the general public have a reasonable understanding of the technique involved and the social and ethical issues that are raised by that methodology. For an explication of that technology and how conclusions are drawn from it, and a sophisticated discussion of the social and ethical problems that arise from its broad application, Genetic Justice is required reading. Sheldon Krimsky is chairman of the board of the Council on Responsible Genetics, and Tania Simoncelli was science adviser to the ACLU and now works for the US Food and Drug Administration.

As Krimsky and Simoncelli explain, sprinkled throughout the genome of humans are short pieces of DNA that are repeated in tandem and have no apparent physiological or developmental function (short tandem repeats, or STRs). The basic chunk is usually four nucleotides in length and is made up of some combination of the different nucleotides, adenine (A), thymine (T), guanine (G), and cytosine (C). There are then 256 possible chunks. One chunk might be, say, ATGG, another GATG, another CATA, etc. In one place on a particular chromosome in one person’s genome there might be three tandemly repeated chunks of a particular kind, say GATGGATGGATG, while someone else might have six repeats of this chunk. Moreover, since everyone has two examples of every chromosome, one inherited from the mother and one from the father, each person has a possibility of different repeat lengths on the two copies of the chromosome. So our hypothetical person could have three repeats on one chromosome and five repeats on the other member of the pair. A different person will have a different number of repeats of the two members of this pair, say twelve and nine.

To identify a person by his or her DNA, forensic scientists have chosen thirteen different STR sites, each on a different one of the twenty-four different chromosomes in the human genome. Each of these sites has its characteristic chunk composition, but there will be immense variation among individuals in any human population in the number of repeated chunks of each of the thirteen kinds. So, for chromosome 7, say, one person might have five repeats on the copy inherited from her father and twelve repeats on the copy inherited from her mother while another person might have inherited three copies from one parent and nine from the other.

The question then is how many people will have the same set of copy number pairs over all thirteen STR sites, since it only takes one difference to rule out someone as having been the source of DNA that has been recovered from a crime scene. Some copy numbers will be very rare in the population and some much more common. The authors give a hypothetical example in which a suspect’s sample has between four and fourteen repeats for different STRs and in which a hypothetical population corresponding roughly to what has been observed in practice has frequencies between .01 and .10 for the thirteen copy number pairs. Multiplying these thirteen frequencies gives the probability of finding by pure chance such an observed DNA type in the hypothetical population. The answer turns out to be about 5 in a billion billion! That is, the suspect’s DNA profile is for practical purposes unique. If such DNA is found at the crime scene, the suspect (or his identical twin) must have been there. Or perhaps not.

Genetic Justice begins with a true story. In 1993 a woman in Germany was strangled to death and some DNA belonging to a woman, not the victim, was found on a teacup. Eight years later, an antiques dealer in another part of Germany was found strangled and the same DNA, not hers, was recovered from items in her shop and the door handle. Then, in 2007, a German policewoman was shot from behind in her patrol car in Heilbronn and again the same DNA was retrieved. After that the same DNA was recovered from numerous crimes of a great variety.

By 2009 “Germany’s most dangerous woman” was wanted for at least forty crimes in Germany, Austria, and France. And then she was found, working in an Austrian factory that manufactured the swabs used by the police to collect DNA at crime scenes. A major problem with DNA as evidence is that we all leave it everywhere that we have been. Cells that are sloughed off our fingers, hairs that we shed, the sneeze that we spray if we do not get to our handkerchiefs in time, the traces of DNA in the saliva with which, old-fashioned as we are, we sealed the envelope of a note to a friend, all contain enough of our DNA to provide a profile to a reasonably competent professional laboratory analyst. It is not sufficient to prove guilt that enough DNA to yield a profile is found somewhere at the scene of a crime. It matters whether the DNA from a rape is found in large quantities on the victim’s underclothes or in her vagina, rather than in trace amounts on her outer garments or other belongings.

Yet another important issue raised by Krimsky and Simoncelli is the problem of technical care and competence, both in collecting samples at crime scenes and in their analysis in police laboratories. There is a great variation in the technical care and competence among various local laboratories. In New York City there is relatively careful quality control. For example, for every DNA crime sample that is analyzed there must be a control run involving a known DNA sequence; and the FBI will not report an analysis based on fewer than ten STR sites. But such ample resources and the scrupulousness of laboratory technicians are not uniform throughout the country. An independent quality control report on the Houston Police Department’s crime laboratory found that among 135 DNA cases reviewed, 43 involved serious analytic errors.

The widespread use of DNA profiles by the police has led, in a number of countries, to the establishment of national computerized data banks to which local departments can contribute and refer. These record the DNA profiles of people arrested, whether or not they have been convicted of a crime, and of DNA collected at crime scenes. In the United States the FBI manages such a database, the Combined DNA Index System (CODIS), which currently contains over eight million records. CODIS, or a similar database in other countries, is used in “dragnets.” If DNA is recovered from a crime scene or taken from anyone remotely connected to a crime, it can be used to search CODIS. If there is a perfect match, a prime suspect has been identified from what is known as a “cold hit.” In case this suspect had already been imprisoned or is dead or otherwise physically removed from any possibility of having committed the crime, one must assume that a laboratory error or a contaminant during the investigation is the source. Putting aside such annoying situations, all other evidence exculpating the suspect is explained away or ignored.

A second form of dragnet does not require perfect matches in CODIS but is based on social presuppositions including racial profiling or on geographic proximity. For example, in a serial rape investigation in Miami, a dragnet by Florida officials asked more than 120 Hispanic males to provide DNA samples, and in Ann Arbor, Michigan, police asked more than 600 black men to submit to DNA testing in another serial rape case, of whom 160 agreed. The latter example exemplifies an important issue in civil liberties and the legality of searches and seizures. The police chief in Ann Arbor threatened that anyone who refused to submit a sample would become an active suspect in the case.

While the probability of finding two unrelated people with the same complete thirteen-site DNA profile is essentially zero, close relatives may have the same profile for part of the thirteen sites. This creates the possibility of “familial searches” for relatives of the actual criminal who might then lead to a final resolution of the case. While only identical twins have identical profiles, full siblings will share, on the average, half of their alleles, as will a parent and offspring, while cousins will share a quarter. Once a group of likely relatives of the guilty person has been identified, the police may then narrow down their search sufficiently to use other information in their search.

Partial-match dragnets and familial searches raise serious issues of civil rights. Both involve taking DNA from the bodies or secretions of persons who have not voluntarily come forward with an offer of their saliva or blood. In many cases a great deal of pressure must be applied in the form of threats against people who have not themselves been charged with any crime, as in the Ann Arbor case, or else there may be surreptitious sampling of objects in their possession. For Krimsky and Simoncelli these are all forms of searches and seizures without legal warrant that are forbidden by the Fourth Amendment. If the police cannot enter my house and search through or seize the physical objects contained in it without a legally obtained warrant, surely my body must be so protected. There is even the question of whether the cells and hair that I shed, the saliva that I deposit on the flap of an envelope, and the feces that I eject are “abandoned” materials and therefore are freely available to investigative authorities (or anyone else who cares to have them). Thus far, the courts have held them to be so, although it is unclear how the judges would proceed against persons who wished to maintain their proprietary rights.

While DNA analysis has become a major tool for the apprehension and conviction of people who have committed acts of violence, it has also led to the exoneration of those who were falsely accused. Krimsky and Simoncelli have both been deeply involved in the struggle to exonerate the innocent. Their book, as might be expected of them, contains chapters on such topics as “Privacy and Genetic Surveillance,” “Racial Disparities in DNA Data Banking,” and “Toward a Vision of Justice: Principles for the Responsible Uses of DNA in Law Enforcement.”

Having read the last chapters of Genetic Justice, one should then turn to Convicting the Innocent by Brandon Garrett, a book devoted to the deliberate use of DNA primarily as a tool for exonerating and freeing those who, for whatever reason, have been falsely convicted and punished. Garrett’s book arises from his longtime association with Barry Scheck and Peter Neufeld, founders and directors of the Innocence Project, whose purpose is to overturn false convictions through the use of DNA evidence. As of this year more than 250 people, convicted of a variety of serious crimes, have been exonerated by the use of DNA, many of them after years in prison and even the threat of execution.

One of the ironies of that history is that Scheck and Neufeld, together with the author of this review, originally opposed the use of DNA evidence. That is because, in its early stages, DNA identification used DNA variants that differed very much among ethnic and racial groups in their relative proportions. An example was the murder, in a trailer park in northern Vermont, of an Abenaki Indian woman. About a third of the residents of the trailer park were Abenaki in a region with many Abenaki. They have proportions of the various DNA variants that are quite different from Caucasians. An accused Abenaki had DNA that matched blood from the crime scene, but the trailer park is on a main north–south tourist route. So whose population frequencies should have been used in making a case?

All that has changed now. The differences in proportions of the DNA variations currently used in forensic analysis make only a trivial difference in conclusions about incrimination. The Innocence Project can exculpate people beyond any reasonable doubt without having to make arguments about ethnic group differences.

Garrett has chosen to illustrate the various sources of injustice in prosecuting and convicting innocent people. These include false confessions extorted by extreme psychological pressure on vulnerable people who are promised relief (as with the Central Park Five, among many others); misidentification by eyewitnesses who are themselves under police pressure; eyewitness identification from carefully stacked lineups and subject to police management; flawed forensic tests that include hair matching, shoeprints, fingerprints, and voice matching; and out-and-out lying by witnesses or jail informants who have something to gain by making things up.

A uniquely valuable part of Garrett’s book is a statistical appendix that provides a quantitative overview of the false convictions, their consequences, and the factors that contributed to them. For example, 47 percent of exonerations occurred between sixteen and thirty-five years after the original convictions. Of those exonerated, 62 percent were black, and 89 percent were accused of rape. In 74 percent of cases forensic evidence was presented against them and 76 percent of convictions were backed up by eyewitnesses. It is hard to imagine seven pages more damaging to the claims of our system of criminal justice, a system in which careerism is deeply embedded.

In a recent issue of The New York Times Magazine a leading state’s attorney for Lake County, near Chicago, when asked by a reporter for The Chicago Tribune about what seemed to be contradictions in a case, is quoted as replying: “The taxpayers don’t pay us for intellectual curiosity. They pay us to get convictions.” I was immediately reminded of Floria Tosca’s words as she contemplated the fallen police chief Scarpia: “Avanti a lui tremava tutta Roma.” Before such prosecutors all America must tremble.

1 For a remarkable essay that places this set of events and the reactions to it in the social and ideological milieu of New York see Joan Didion’s discussion in ” New York: Sentimental Journeys ,” The New York Review , January 17, 1991. Sarah Burns does not mention this brilliant article, which subjected evidence to skeptical analysis and anticipated later revelations. ↩

2 Erica Goode, “When DNA Evidence Suggests ‘Innocent,’ Some Prosecutors Cling to ‘Maybe,'” The New York Times , November 16, 2011. ↩