May 2003, Page 24

Evaluating Forensic DNA Evidence, Part 2
By William C. Thompson; Simon Ford; Travis E. Doom; Michael L. Raymer; Dan E. Krane


Breaking open the black box: How to review the electronic data

Reviewing the electronic files produced by the ABI Prism 310 Genetic Analyzer™ (or similar equipment) has a number of additional benefits beyond revealing unreported low-level peaks. The software that controls these devices creates a complete record of all operations the device performs while typing samples in a particular case and records the results for each sample.

These records can reveal a variety of problems in testing that a forensic laboratory may fail to notice or choose not to report, such as failure of experimental controls, multiple testing of samples with inconsistent results, re-labeling of samples which can flag potential sample mix-ups and failure to follow proper procedures. We know of several cases in which review of electronic data has revealed that the laboratory failed to run all of the necessary control samples needed to verify the reliability of the test results, or that the laboratory ran the control samples under different conditions than the analytical samples (a major breach of good scientific practice).

The electronic files are also useful for producing trial exhibits. An expert with the right software can convert the files from their proprietary format into Adobe Acrobat™ files containing images that can easily be inserted into Powerpoint™ and Microsoft Word™ documents.

It is easy for crime laboratories to produce the electronic data that underlie their conclusions. All that is necessary is to copy the files produced in the case onto a CD-ROM or other storage medium. CD-ROMs are generally preferred because they create an unalterable record of the data produced by the laboratory. Copying files to a CD-ROM is a simple point and click operation that can be accomplished in fifteen minutes or less in most cases. CD-ROM burners compatible with any laboratory computer are available commercially for under $200. There is no legitimate excuse for refusing to turn over electronic data for defense review. In a few instances laboratories have resisted producing electronic files, or have even destroyed the files, but the great majority of trial courts will not tolerate such obstructive behavior.

The electronic data produced by the ABI 310 Genetic Analyzer™ is in a proprietary format that can only be read and interpreted by ABI’s Genescan™ and Genotyper™ software. Defense lawyers seeking a review of electronic data must find an expert who has access to this software. The review process typically takes a minimum of 3-4 hours, and may take much longer in an even moderately complicated case. Another option is provided by a company called Forensic Bioinformatics (www.bioforensics.com), which can analyze electronic data using an “expert system” software tool called Genophiler™. Genophiler, which was developed by the authors of this article, automates Genescan™ and Genotyper™ analysis, presents convenient summaries of the analyzed data, and identifies and flags many of the technical issues discussed in this article.


Are there innocent explanations for the lab’s findings?

In many cases, careful review of the underlying laboratory notes, electropherograms and electronic data will reveal no significant problems. Defense lawyers should never forget, however, that even clear-cut DNA test results may have innocent explanations.
Sample handling errors. Accidental mix-up or mislabeling of samples is a possibility that always must be considered. We have encountered a number such errors while reviewing case work.1 In most instances the mix-ups readily come to light (and are caught by the lab) because they produce unexpected results: Samples that are supposed to be from a man show a female DNA profile, two samples known to be from the same person show different DNA profiles, and so on. The real danger arises when sample mix-ups produce plausible results. In these instances, forensic analysts may overlook subtle clues that something is amiss because they expected to find the very result produced by their error.

For example, after reviewing the laboratory notes in a Philadelphia rape case, one of the authors noticed some clues (later confirmed by additional testing) that the Philadelphia Police Crime Laboratory had mixed up the reference samples of the defendant and the rape victim. This mix up had falsely incriminated the defendant because the lab found what it thought was the defendant’s DNA profile in a vaginal swab from the victim. In fact, it was the victim’s own profile, and was mistakenly matched to the defendant due to the mix up.2 Similar errors have come to light in other cases. Cellmark Diagnostics mistakenly mixed up the victim and defendant in a San Diego rape case, thereby mistakenly incriminating the defendant.3 

A similar error occurred in Las Vegas, where a mix up involving reference samples of two men sent one man to jail for a rape committed by the other. The innocent man had been incarcerated for over a year when the error came to light in April, 2002.4 In both cases the error came too light only after a defense expert noticed inconsistencies in the laboratory records.

It is not always possible to tell from the laboratory records whether samples actually were mixed up or cross-contaminated. However, careful review of the laboratory records will usually provide important information about whether such errors could have happened. For example, evidence that a reference sample from the defendant was handled or processed in close proximity to samples from the crime scene can support the theory that a sample handling error explains incriminating results. In one case, review of a criminalist’s notes showed that the defendant’s trousers, collected at his home, were transported to the laboratory in the same box that contained a number of items from the crime scene that were saturated with the victim’s blood. This fact cast important new light on a seemingly incriminating result: blood from victim was detected on the defendant’s trousers.

We suggest that defense lawyers obtain and review complete copies of all records related to evidentiary samples collected in the case (see Appendix for a model discovery request). It should be possible to document the complete history of every sample from the time it was initially collected through its ultimate disposition.


Inadvertent transfer of DNA
One of the most striking developments in forensic DNA testing in recent years is the testing of ever smaller biological samples. Whereas the original DNA tests required a fairly large amount of biological materials to get a result (e.g., a blood stain the size of a dime), current DNA tests are so sensitive that they can type the DNA found in samples containing only a few cells. There is likely to be enough of your DNA on the magazine you are reading right now for your DNA profile to be determined by a crime lab.

The increasing sensitivity of DNA tests has affected the nature of criminal investigations and has created a new class of DNA evidence. Analysts talk of detecting “trace DNA,” such as the minute quantities of DNA transferred through skin contact. DNA typing is currently being applied, with varying degrees of success, to samples such as doorbells pressed in home invasion cases, eyeglasses found at a crime scene, handles of knives and other weapons, soda straws, and even single fingerprints.

These developments will bring more DNA evidence to court in a wider variety of cases and may well open new lines of defense. A key issue will be the potential for inadvertent transfer of small amounts of DNA from one item to another, a process that could easily incriminate an innocent person. Studies have documented the presence of typeable quantities of human DNA on doorknobs, coffee cups and other common items.5 

Studies have also documented the inadvertent transfer of human DNA from one item to another.6 Primary transfer occurs when DNA is transferred from a person to an item. Secondary transfer is when the DNA deposited on one item is transferred to a second item. Tertiary transfer is when the DNA on the second item is, in turn, transferred to a third. There are published studies that document secondary transfer of DNA (in quantities that can be detected by STR tests) from items that people simply touched to other items.
A recent study commissioned by a wealthy defendant was used to show that tertiary transfer of DNA could have occurred in a manner that falsely incriminated the defendant. Dr. Dirk Greineder, a prominent physician and adjunct Harvard professor, was accused of killing his wife. A DNA profile similar to Greineder’s was found, mixed with his wife’s profile, on gloves and a knife found near the crime scene. Greineder denied touching these items, which appeared to have been used by the killer. But how did his DNA get on them?
Greineder offered a two-pronged defense.

First, he challenged the conclusion that his DNA matched that on the gloves, noting inconsistencies between his profile and the profile on the gloves. The crime laboratory had shifted its threshold for scoring alleles in a manner that allowed it to count alleles that matched with Greineder, while ignoring some that did not. And the lab had to evoke the theory of “allelic drop out” to explain why some of Greineder’s alleles were not found.

Greineder’s second line of defense is our focus here. He argued that his DNA could have gotten onto the glove through tertiary transfer. He and his wife had shared a towel the morning of the murder — perhaps his DNA was transferred from his face to the towel, and from the towel to his wife’s face. His wife was later attacked by a glove-wearing stranger who struck her on the face, strangled her, and stabbed her, in the process transferring Greineder’s DNA from his wife’s face to the gloves and the knife. According to this theory, the tell-tale extra alleles on the gloves and knife that matched neither Greineder nor his wife were those of the killer.

To support the theory that his DNA could have been transferred innocently to the instruments of murder, Greineder commissioned a study. Forensic scientists Marc Taylor and Elizabeth Johnson, of Technical Associates (an independent laboratory in Ventura, California) simulated the sequence of events posited by the defense theory: A man wiped his face with a towel, then a woman wiped her face with the towel, then gloves and a knife like those used in the murder were rubbed against the woman’s face. DNA tests on the gloves and knife revealed a mixture of DNA from the man and woman — exactly what was found in the Greineder case.7 Taylor was allowed to present his findings to the jury. Although the jury ultimately convicted Greineder (there was other incriminating evidence besides the DNA) the case is a good example of how the amazing sensitivity of contemporary DNA profiling methods facilitate a plausible explanation for what might at first seem to be a damning DNA test result.


Finding experts

The complexity of short tandem repeat (STR) testing makes it difficult if not impossible for a lawyer to evaluate the evidence without expert assistance. Defense lawyers generally need expert assistance to look behind the laboratory report and evaluate whether its conclusions are fully supported by the underlying data. Defense lawyers may also need expert assistance to develop and assess alternative theories of the evidence. Experts can also be helpful, and often are necessary, to assess whether laboratory error or inadvertent transfer of DNA might plausibly account for the incriminating results.

In our experience, the best experts for evaluating whether the lab’s finding are supported by the underlying data are academic scientists in the fields of molecular biology, biochemistry, bio-informatics, molecular evolution, genetics (particularly human and population genetics), and related fields. It is not essential that the expert have had experience analyzing forensic samples. In fact, we find that forensic scientists often (but not always) make poor defense experts because they tend to accept too readily the goal-directed subjective judgments and circular reasoning of their crime lab colleagues.

Academic scientists generally have much stronger training in scientific methods and, as a result, demand that test results be interpreted in a scientifically rigorous and unbiased manner. They often are appalled at the willingness of some forensic scientists to rely on subjective judgment and guesswork to resolve ambiguities in scientific data and their unwillingness to utilize blind procedures when making such judgments.

Having the electronic data analyzed by a company like Forensic Bioinformatics (www.bioforensics.com) is a good first step and can make it easier to work with an expert. Such automated analyses eliminate the need for the expert to do several hours of tedious work that requires specialized software, making it possible for the expert to get to the heart of the matter more quickly. They also highlight potential issues and problems that the attorney can use to get the interest of an expert.


Conclusions

Careful review of DNA evidence can reveal a variety of potential weaknesses, making it possible in some cases to challenge the government’s conclusions and offer alternative interpretations. In order to provide effective representation to a client incriminated by DNA evidence, the defense attorney must do more than simply read the laboratory’s conclusions. It is important to obtain and review the underlying scientific records, including electronic data, in order to determine whether the laboratory’s conclusions are fully supported by the test results. It is also important to evaluate alternative explanations for the test results, to determine whether there are plausible innocent explanations. Promoters of DNA testing have effectively used the media to convince most people, including potential jurors, that the tests are virtually infallible. As DNA testing becomes more common in the justice system, it is vital that defense lawyers give it careful scrutiny in order to detect and expose those cases where genetic evidence deserves less weight than it is otherwise likely to receive.



Notes

1. See, William C. Thompson, Franco Taroni, and Colin G. Aitken, How the probability of a false positive affects the value of DNA evidence, J. Forensic Sci. (January 2003, in press).
2. See Id. for further discussion of this case. Copies of the laboratory reports may be obtained from William C. Thompson.
3. Id.
4. Glen Puit, DNA Evidence: Officials admit error, dismiss case. LV lab put wrong name on sample, Las Vegas Review-Journal, April 18, 2002.
5. See, van Oorschot DNA fingerprints from fingerprints, Nature, June 19,1997, 767; Findlay, et al., DNA fingerprinting from single


National Association of Criminal Defense Lawyers (NACDL)
1150 18th St., NW, Suite 950, Washington, DC 20036
(202) 872-8600 • Fax (202) 872-8690 • assist@nacdl.org