5

Learning from Abroad

In this workshop session, moderated by Natalie Ram, University of Maryland King Carey School of Law, presenters provided perspective on law enforcement use of emerging forensic DNA technologies abroad. Presenters included individuals from Australia, Portugal, the United Kingdom, and Switzerland with expertise in forensic genetics, biochemical engineering, sociology, law, criminal justice, and forensic medicine and molecular biology. Their presentations highlighted the diverse approaches to regulation and use of advanced forensic DNA technologies by law enforcement across the United Kingdom, Europe, and Australia.

BEST PRACTICES AMID A REGULATORY VACUUM IN AUSTRALIA

Dennis McNevin, University of Technology Sydney, began the session by describing the status of each of the advanced forensic DNA technologies in Australia. He started with probabilistic genotyping (PG), noting that many speakers had emphasized the importance of testing PG for accuracy and reliability. McNevin said that he is not aware of any interlaboratory comparisons of PG results among forensic laboratories in Australia; one lab participated in an external proficiency test, but the results are not publicly available. He noted that developers of PG have argued against interlaboratory comparisons because of differences in human factors, lab policies, and other elements extrinsic to the software. McNevin said that it is reasonable

to expect that results should be comparable across labs, and that comparisons can account for interlaboratory differences (McNevin et al., 2021).¹

Forensic DNA phenotyping (FDP) for externally visible characteristics is used in Australia, said McNevin, as is testing for inference of biographical ancestry (BGA). He said that in his view, BGA should not be considered a type of FDP. He explained that phenotyping refers to an essential quality of an individual as a result of their DNA, whereas ancestry estimation indicates where an individual’s ancestors may have come from but does not reveal any “essential characteristic” of the individual. McNevin noted that Australian law enforcement currently uses FDP only to infer eye and hair color, and not to infer skin color or other externally visible traits. McNevin noted that in Australia, FDP and BGA are typically employed only if other testing—such as short tandem repeat (STR) direct matching and familial searching—fails to confirm an identity. In addition, privacy impact assessments are now typically conducted before implementing FDP or BGA (Box 5-1). McNevin said that in his opinion, phenotypes and ancestry inferences should be retained in house for investigative purposes and not released to the public in order to avoid possible stigmatization of certain communities. Describing the legal context, McNevin explained that there is a legislative vacuum for both FDP and BGA in Australia; neither is explicitly allowed or disallowed.

There is also a legislative vacuum around forensic investigative genetic genealogy (FIGG), said McNevin. As a result, FIGG operates under a privacy umbrella (e.g., Commonwealth Privacy Act, 1988, and amendments). Under these regulations, FIGG is subject to regulatory aspects designed to ensure ethical and transparent use (Box 5-2). McNevin also noted that the use of privacy impact assessments is widespread in Australian law enforcement and is an integral part of FIGG implementation.

McNevin explained the process of implementing FIGG in cases of missing persons as part of a framework of sequential unmasking of information. Cases begin with testing autosomal STRs and uploading the profile to the National Criminal Identification DNA Database (NCIDD). If there is no hit, the profile would be uploaded to the familial searching extension of NCIDD, and next a Y-chromosome STR (Y-STR) or mitochondrial DNA profile could be uploaded to the same database to look for maternal or paternal relatives. If there are no meaningful leads after these steps, the STR profile could be uploaded to the International Criminal Police Organization

___________________

¹ As McNevin described, this comparison method involves applying PG to a dilution series of a DNA mixture to determine where the likelihood ratio plateaus. Because different laboratories use different assays with different loci, this method includes only loci that are common among all participating labs. In addition, each lab uses the same population allele frequencies and subpopulation correction factor theta. With this method, said McNevin, interlaboratory comparisons are possible.

(INTERPOL) DNA database, and then to I-FAMILIA (INTERPOL Family Associated Matching to Identify Lost Individuals Abroad). Following this, Y-STRs and mitochondrial DNA could be searched in I-FAMILIA, and then Y-STRs would be searched using the Y-STR Haplotype Reference Database and mitochondrial DNA to the EMPOP Database.² Finally, said McNevin, if all of these steps failed to generate a meaningful lead, a single nucleotide

___________________

² EMPOP stands for the European DNA Profiling Group mtDNA Population (Database).

polymorphism (SNP) profile would be generated in order to undertake the process of FIGG (see Figure 5-1).

It is essential, said McNevin, to educate and “bring the public along” in the use of FIGG. The Australian Federal Police DNA Program for Unidentified and Missing Persons engages with the public through its website, which provides information to families of missing persons as well as members of the public who may wish to contribute DNA to assist in identifying missing persons.

TENTATIVE USE OF FIGG IN EUROPE

Rafaela Granja, University of Minho, focused her remarks on the use of FIGG across Europe and the unique nature of the professional origins of FIGG. She reported growing interest in FIGG in Europe, although it is still in very early stages. Granja explained that Sweden recently used FIGG for the first time to solve a double-murder cold case that occurred in 2004. Since Swedish law does not specifically allow the use of genetic information for FIGG, the Swedish Authority for Privacy Protection has blocked further use of the technology until the law is changed. Granja explained that the Biometrics and Forensics Ethics Group in the United Kingdom stated that the preliminary use of FIGG in the identification of unidentifiable human bodies would allow its potential to be tested. However, she continued, the agency has raised concerns about the adequacy of traditional informed

consent in the context of FIGG and has highlighted the need to reconsider traditional notions of informed consent to ensure ethical and responsible use of genetic data in forensic investigations. This concern about traditional conceptions of informed consent echoed concerns expressed during the workshop discussion of PG.

While FIGG gained traction after the Golden State Killer investigation, said Granja, the story of FIGG has longer roots. People have been engaging in genetic genealogy as a hobby for many years, and this field of expertise has emerged as a form of “citizen science.” Early on, this often took the form of law enforcement agencies and citizen genealogists working together to solve missing persons cases. This partnership, she said, means that hobbyists with diverse backgrounds have become a source of expertise for criminal investigations. It is “of paramount importance” that this community’s involvement with the criminal justice system is formalized and that safeguards are put in place. These safeguards could include accreditation or regulatory procedures that limit who can play a role in this system. Current safeguards, she said, are largely self-regulatory efforts; instead of self-regulation, there is a need for a diverse group of stakeholders to make decisions about the complex issues involved with FIGG. FIGG has largely been shielded from public criticism by the perception that it is used to increase public safety and provide healing for victims, said Granja. However, law enforcement use of recreational DNA databases is an example of “function creep,” with implications for civil liberties. FIGG and other forensic DNA technologies are social issues with social implications, she said, and privacy needs to be considered a social value.

A LACK OF CLARITY IN ENGLAND AND WALES

Carole McCartney, University of Leicester Law School, told workshop participants about the use of forensic DNA technologies in England and Wales. It is an entirely private forensic science marketplace, she said, with the police contracting with one of the three large providers. The commercial nature of the marketplace for DNA testing means that the contracts between the police and the providers dictate which technologies are available and complicate the issue of transparency. It is not a healthy marketplace, she said; there is enormous financial pressure on the companies to sell their technologies, and prices have been forced down by the risk of company failure when providers must compete, as well as by cuts to police budgets. The government has not stepped in with a strategic plan for advanced forensic DNA technologies, said McCartney; the last U.K. Home Office (2018) document, Biometrics Strategy: Better Public Services Maintaining Public Trust, contained a paucity of detail and no underlying principles of future strategy on forensic DNA.

In gauging the current use of advanced DNA technologies in England and Wales, stakeholders and users provided conflicting reports. While stakeholders told McCartney that PG is used routinely, some stakeholders suggested that only STR testing is allowed by law enforcement in most investigations and that FDP and FIGG may be restricted; others reported that there were increasing instances of these technologies being utilized or piloted.

This “general confusion” around the use of forensic DNA technologies reinforces an ambiguous legal context, she said. Only two laws—the Police and Criminal Evidence Act (1984) and the Data Protection Act (2018)—have anything to say about the use of retained DNA profiles. McCartney explained that these laws have broad parameters and a lack of detail to guide potential users as to DNA use in law enforcement. For example, the Police and Criminal Evidence Act (1984) states that retained DNA must not be used other than “for purposes related to the prevention or detection of crime, the investigation of an offence or the conduct of a prosecution (or the identification of the deceased)” (c. 60). McCartney noted that if evidence from advanced forensic DNA technologies was to be used in court, the Criminal Procedure Rules and the Criminal Practice Directions would apply; she noted that these rules would be “highly problematic” for the admission of this DNA evidence.

A judicial primer on forensic DNA analysis is designed to inform courts for making decisions related to DNA evidence, but McCartney noted that its guidance is outdated and it does not include these technologies. The Forensic Information Database Board determines who can access the National DNA Database, said McCartney. They have yet to say anything about these new technologies being used but have previously agreed to the sending of samples overseas for such testing. Other entities—such as the Biometrics and Forensics Ethics group, the Forensic Regulator, the Biometrics Commissioner, and the Information Commissioner—look at individual aspects of forensic DNA technologies (e.g., privacy) but do not provide a wider or strategic, future-looking view.

When considering how to move forward, said McCartney, regulation of advanced forensic DNA technologies could take the form of law or code(s) of practice. Both options are complicated, she said. Laws are difficult to write well and would require determining what department would oversee advanced forensic DNA technologies, what the power and resources would be for oversight, and what legislative vehicle would be used. Relying on codes of practice can give the appearance of a practice being regulated when it is not, and determining what is “best practice” is rarely straightforward.

In conclusion, McCartney said oversight of advanced forensic DNA technologies in England and Wales is characterized by an overreliance on a disparate collection of policies, guidance, and codes of practice that are

not publicly debated or publicly available. Current laws are confusing and overly broad, and there is no consensus on “best practice.” Technological advances attract the attention of ethicists but there is far less attention paid to the regulation of these advances. She quoted the biometrics commissioner who said in 2018 that “actual deployment of new biometric technologies may lead to more legal challenges unless Parliament provides a clear, specific legal framework for the police use of new biometrics.” Science is “marching on while law and regulation are still dragging their heels,” said McCartney.

MULTIPLE LEVELS OF REGULATION IN SWITZERLAND

Martin Zieger, University of Bern, described the regulatory scheme in Switzerland around forensic DNA technologies. There are three levels of relevant regulation: the Criminal Procedure Code, the DNA Profiles Act (2003), and two technical ordinances. Together, these regulations determine whether and how advanced forensic DNA technologies can be used. Zieger reported that the DNA Profiles Act was recently revised to permit FDP (Revision of the DNA Profiles Act, 2023); this revision took 8 years from start to finish. Ziegler stated his understanding that FDP is currently restricted to serious offenses against physical and sexual integrity, and the use of FDP must be ordered by the public prosecutor, rather than the police. After 5 years of use of FDP, an evaluation will be conducted to see whether it has provided useful investigative leads, Zieger explained. He noted that the law specifically allows the ascertainment of eye, hair, and skin color; biogeographical origin; and age, and it specifically prohibits evaluating health-related or personal characteristics such as character, behavior, or intelligence. Zieger explained that when writing this law, parliament added a provision that allows the government to add permitted characteristics if “practical reliability is given.” He said that while this was intended to allow the government to respond quickly to changes in technology without waiting for Parliament, it opens the door to an undemocratic process in which the executive power, based on recommendations from the federal police department, could make unilateral decisions. FDP has been permitted in Switzerland since August 2023, but Zieger said he knows of only one case that has used it.

FIGG is not regulated in Switzerland, so it is widely deemed inadmissible in court and is largely absent from the public debate. Zieger said that when the DNA Profiles Act was under revision (2023), only 2 out of 51 interested parties said that regulating FIGG was a priority. A national human genetic testing law has “harsh standards” for direct-to-consumer tests, which in theory makes it difficult for police to access consumer DNA databases. Switzerland law enforcement does engage in familial searching, but it

is not widely used; he said that in the last 8 years there have been 20 cases in which familial searching was used, and only 1 in which it was successful.

PG is not mentioned in the DNA Profiles Act (2003), but all laboratory procedures are subject to control by the Swiss Accreditation Service. These standards require labs to participate in proficiency tests for any calculation software used in the lab. Zieger said that to his knowledge, two out of seven labs in Switzerland are using PG and one only uses it for operational casework. He hypothesized that the low uptake is because “most biologists are not big fans of statistics.”

In closing, Zieger made three recommendations. First, he suggested that mass DNA tests should not be permitted based on FDP alone. Second, he called for rules for data access and the handling of incidental findings when conducting FDP or FIGG. Finally, he said that software validation and proper training are key for laboratories conducting PG.

GENERAL DATA PROTECTION REGULATION

A workshop audience member asked the panelists to comment on the General Data Protection Regulation (GDPR) and whether it impacts the use of forensic DNA technologies. McCartney replied that laws like the GDPR are often quite vague. There are provisions that say that the use of genetic data is prohibited, but a later provision provides a “massive get out clause” by allowing the use of genetic data if it is for law enforcement. Whether and how the law impacts the use of forensic DNA technologies depends on who is interpreting and overseeing the law, and whether guidance is provided about specific applications. Zieger added that GDPR explicitly excludes law enforcement use from the privacy law, but it does not give clear guidance on any limits of this exclusion. In his view, GDPR does not apply to the use of forensic DNA technologies for law enforcement. Instead, forensic DNA technologies are subject to the Law Enforcement Directive of the European Union, which leaves much leeway for implementation to the member states. Granja said that GDPR does impact FIGG because it explicitly requires users to give consent for their DNA to be used for law enforcement purposes.

BLIND TESTING

Following up on McNevin’s mention of sequential unmasking of information, Norah Rudin, Forensic DNA Consulting, commented during the Q&A that she was part of an interdisciplinary group that introduced the process called sequential unmasking into the forensic space for use in casework. Blind testing, though standard in clinical testing, is generally not typical in forensic science, she said. As a compromise, her group came up with the idea of sequentially introducing information that could be relevant to

the analysis of crime samples while staying as blind as possible for as long as possible. It ensures that labs receive the necessary data for a comprehensive and informed analysis while documenting how each piece of information is introduced and its impact on the conclusions drawn. Zieger added that, regarding proficiency testing, blind testing in the laboratory would mean receiving traces and not knowing that they are part of a test; he said this could be difficult to achieve but would be “really interesting” because it introduces some bias in the evaluation, if the analyst knows that they treat a test and not a real casework sample. Rudin replied that the method she and her colleagues developed for use in casework suggested using firewalls in the lab so that one person is communicating with law enforcement to obtain information about which samples are relevant to the questions in the case. The person doing the analytical work does not initially know, for example, the identity of the samples or what is the question in the case; these pieces of information are revealed as necessary going forward. Alicia Carriquiry, Iowa State University, added that while blind testing is challenging, it is not impossible. The Houston Forensic Science Center has been implementing blind testing in several disciplines for many years, she said, and there is an effort in Texas to implement blind testing across crime labs.