Saturday, January 18, 2020

A Fourth Model of Law Enforcement Access to DNA Databases

First there were DNA databases of four or five RFLP VNTR measurements obtained by gel electrophoresis of the DNA fragments of convicted offenders. These were soon superseded by much larger databases for convicted-offenders (and later, arrestees) comprised of a larger number of STRs determined with capillary electrophoresis. The individuals supplying the DNA analyzed in these limited ways had no choice in the matter. Statutes compelled them to submit to DNA sampling, and only law enforcement authorities had access to these special-purpose, government-run databases.

In recent years, the potential for private databases to generate investigative leads through trawling (without individualized suspicion or probable cause) has begun to be exploited. Millions of SNP-based records reside on the servers of recreational genetics companies that process saliva samples with SNP arrays that determine the alleles present at hundreds of thousands of SNP loci. The customers of each direct-to-consumer (DTC) testing company can learn if other participating customers of that company have large blocks of DNA in common -- a situation indicative of a family relationship. (Other tests for relatedness in the private databases are also possible, but the haploblock-matching procedure has been the most productive for criminal investigations because it casts a wider net.) In addition, over a million samples reside in the voluntary private database known as GEDmatch, which enables genealogy enthusiasts to share their personal SNP-array data.

SNP-array testing does not work well with crime-scene DNA samples that contain mixtures of DNA from several individuals, that are severely limited in quantity, or that are degraded. However, an alternative technology, known as massively parallel, or next generation sequencing (MPS or NGS), can generate genomic data from challenging samples, and the sequence data can be culled to supply the SNP-array data that DTC companies would have provided had the source of the DNA evidence from the crime-scene sent a saliva sample in to one of these companies. Armed with such a data file, police who are able to access a DTC or GEDmatch-type database can look for potential relatives. By turning to these individuals (or public records about the families), police sometimes can find a suspect who merits further investigation.

Many DTC customers who did not realize their DNA might lead police to their relatives (known or unknown to them) have found this kind of forensic genomic genealogy (FGG) profoundly disturbing, at least when it was not something they had explicitly signed up for. Others have evinced less concern. The private database operators have responded with different policies. GEDmatch presumes that users object, but allows them to indicate their willingness to have their data used in law-enforcement trawls. Another possible response is an opt-out policy, in which the presumption is that individuals are not opposed to this use of their data. Yet another policy is that the database is always open to law enforcement, just as it is to curious individuals. Finally, the database could be completely closed to law-enforcement trawling (without judicial approval based on a showing of individualized cause).

In short, there are three main models of law enforcement access: (1) the government-run, law-enforcement-only database; (2) the private other-purpose, opt-in database; and (3) the private other-purpose, opt-out database. However, a fourth model is emerging -- a privately run law-enforcement-only database.

An article from the information service GenomeWeb reports on the plans of Othram, which bills itself as "the first technology company to apply all the power of modern sequencing and genomics to forensics" so as to secure "justice through genomics." The article includes a number of interesting statements from the company's CEO, David Mittelman. GenomeWeb explains that
Othram ... introduced DNASolves.com to solicit users of consumer genomics services to upload their data for the expressed desire to help law enforcement solve cold cases.

"Family Tree DNA is doing the opt-out model [with regards to law enforcement], GEDmatch is doing opt-in," said ... Mittelman. "I thought there should be another model," he said. "Since we do nothing but law enforcement, there is nothing to opt out of."

Mittelman, a former CSO at Family Tree DNA parent Gene by Gene ... said "I have enjoyed the consumer genetics and genealogy side, and I certainly enjoyed the medical side, but I saw the forensics market as a market that was underserved and could benefit from the technology that has been widely used and embedded in consumer and medical testing," he said. "It made sense to bring that technology over."

Mittelman credited the developments in the market with both the success of consumer genomics as well as advancements in next-generation sequencing technology. By some estimates, 30 million people have taken an array-based consumer test to date. Meantime, the drop in the price of sequencing, plus ongoing innovation in the field, means that it is now possible to perform whole-genome sequencing of highly degraded samples from crime scenes and then scour large databases to find genetic relatives, constructing genealogies to identify victims or perpetrators.
...
"In 2019, sequencing failed to penetrate the consumer market," said Mittelman. "But forensics is an interesting market where sequencing is superior to arrays," he said. "It is not just that sequencing gives you more information, in a lot of cases it's the only way to get information," he added.
Othram's approach is called Forensic Grade Genome Sequencing. According to Mittelman, array technology has a "high failure rate" when it comes to forensic samples, making sequencing the go-to technology when it comes to working with these kinds of samples.

"You really need special methods," said Mittelman. "We have developed proprietary methods to adapt the worst kinds of DNA to sequencing," he added. "I think in the long term forensics will only work with sequencing, while for consumer, arrays are good enough."

Othram has not yet published on its techniques, but eventually aims to do so, Mittelman said. While the company hones its sequencing capabilities, it is also hoping more customers of consumer services will be moved to upload their data to DNASolves.com. He noted that only a small percentage of those tested have elected to upload data to GEDmatch, meaning the potential exists to grow a new database of a different set of users interested in helping law enforcement.

"Rather than target a small number of people who are genealogy power users, and ask them to help solve crime instead, it seemed to me that you could approach the 30 million who have tested and tell them if you have tested and feel like getting involved, this is how you do it," said Mittelman. "You don't have to be a power user in genealogy to make a difference in crime solving."
Whether DNAsolves.com will attain the critical mass to be a useful investigative database is not guaranteed, so it is not clear that the data donors will be "making a difference in crime solving." Buut even if the database remains small, Othram can market its FGG service as offering the police agency access to a exclusive database. In a seeming excess of enthusiasm about finding "the most distantly related individuals," the website states that
We are all genetically related to one-another [sic]. Genetic genealogy uses DNA information in combination with genealogical and historical records to establish relationships between even the most distantly related individuals. This is a substantial improvement in human identification capability from current forensic testing methods that enable exact or near exact matches. When you contribute your DNA data, you help identify victims, missing persons, and perpetrators of crimes — even if you are a distant genetic relative.

REFERENCE

Justin Petrone, Forensic Genomics Market Advances Due to Consumer Databases, Technology Innovation, genomeweb, Jan. 9, 2020, https://www.genomeweb.com/sequencing/forensic-genomics-market-advances-due-consumer-databases-technology-innovation

Friday, January 17, 2020

What Are the Law Enforcement Implications of the GEDmatch Buyout?

GEDmatch is a free genetic genealogy database that allows people to upload their genomic data from direct-to-consumer (DTC) testing services such as 23andMe, Ancestry, MyHeritage, and Family Tree DNA. It permits cross-company searches for possible relatives among those who elect to participate. To date, roughly 1.3 million people have uploaded their SNP array data to the service, and GEDmatch continues to add about 1,000 people daily. 1/ Scientists and genealogists working for law enforcement agencies have been able to locate the source of DNA evidence from unsolved crimes by discovering a possible relative who supplied his or her data to public kinship searching in GEDmatch. The result has been fluctuating policies with respect to police use of GEDmatch's haploblock-matching software and uploaded data to produce some leads. 2/

Last month (19 Dec, 2019), GEDmatch's founder sent the following email to its registered users:
To GEDmatch users,

As you may know, on December 9 we shared the news that GEDmatch has been purchased by Verogen, Inc., a forensic genomics company whose focus is human ID. This sale took place only because I know it is a big step forward for GEDmatch, its users, and the genetic genealogical community. Since the announcement, there has been speculation about a number of things, much of it unfounded.

There has been concern that law enforcement will have greater access to GEDmatch user information. The opposite is true. Verogen has firmly and repeatedly stated that it will fight all unauthorized law enforcement use and any warrants that may be issued. This is a stronger position than GEDmatch was previously able to implement.

...It has been reported on social media that there is a mass exodus of kits from the GEDmatch database. There has been a temporary drop in the database size only because privacy policies in place in the various countries where our users reside require citizens to specifically approve the transfer of their data to Verogen. As users grant permission, that data will again be visible on the site. We are proactively reaching out to these users to encourage them to consent to the transfer.

... Verogen recognizes that law enforcement use of genetic genealogy is here to stay and is in a better position to prevent abuses and protect privacy than GEDmatch ever could have done on its own.

Bottom line: I am thrilled that the ideal company has purchased GEDmatch. The baby I created will now mature for the benefit of all involved. If anyone has any doubts, I may be reached at gedmatch@gmail.com. I will do my best to personally respond to all concerns.

Curtis Rogers
GEDmatch
Verogen's acquisition has produced angst among observers who worry that the company will exploit it for police purposes. The observation that Verogen "caters to law enforcement" 3/ has become a media meme. Verogen is a 2017 spin-off from Illumina, Inc., the San Diego Company that pioneered and acquisitioned its way to cheap DNA sequencing machinery for research and clinical applications. Illumina also sells the arrays that the DTC companies use, but that technology is not suitable for most crime-scene samples. Verogen's website proudly announces that "Verogen serves those who pursue the truth" by being "the world’s first sequencing company solely dedicated to forensic science. ... Powered by Illumina technology and free of legacy method allegiance, we are uniquely positioned to support forensic labs with innovative solutions purpose-built for the challenges of DNA identification."

The purpose-built solutions free of legacy method allegiance that Verogen now markets use Illumina's technology for massively parallel sequencing by synthesis that culminates in traditional autosomal forensic STR data and much more genomic data.It is superior to capillary electrophoresis, especially for small, degraded, and mixed DNA samples. Competitors market other packages of MPS devices, reagents, and software to forensic laboratories. 4/

How Verogen's ownership will affect police access to GEDmatch is not clear. No doubt, Verogen would like the police to use the database, but it cannot afford to alienate the genealogy enthusiasts who send in their SNP array data. It can provide materials and software that would help forensic laboratories who use sequencing technology to generate the SNP data in the format for trawling GEDmatch for haploblock matches. Indeed, Verogen has a "forensic genetic genealogy product in development for forensic laboratories [that] will actually provide more privacy protection for users of GEDmatch, as the test is focused on kinship analysis for forensic purposes." 5/ But I am curious as to how this purpose-built product will provide more privacy protection than occurs with kinship matching for private purposes.

NOTES
  1. Justin Petrone, Forensic Genomics Market Advances Due to Consumer Databases, genomeweb, Jan 9, 2020, https://www.genomeweb.com/sequencing/forensic-genomics-market-advances-due-consumer-databases-technology-innovation.
  2. Police Genetic Genealogy at GEDmatch: Is Opt-in the Best Policy?, Forensic Sci., Stat. & L., Sept. 21, 2019, https://for-sci-law.blogspot.com/2019/09/police-genetic-genealogy-at-gedmatch-is.html.
  3. Heather Murphy,  What You’re Unwrapping When You Get a DNA Test for Christmas, N.Y. Times, Dec. 22, 2019 ("The new owner, Verogen, said that it would actively fight future search warrants and that users can still opt out of helping police. But Verogen is also a company that has built its business, so far, on catering to law enforcement.").
  4. Brigitte Bruijns, Roald Tiggelaar & Han Gardeniers, Massively Parallel Sequencing Techniques for Forensics: A Review, 39 Electrophoresis 2642-2654 (2018), https://doi.org/10.1002/elps.201800082.
  5. Petrone, supra note 1.

Wednesday, January 15, 2020

OSAC-approved Testimony

According to the charter of the Department of Commerce's Organization of Scientific Area Committees for Forensic Science, better known as OSAC (but not to be confused with the Department of State's Overseas Security Advisory Council, also known as OSAC), "[t]he aims of the OSAC are to:
  • populate the OSAC Registry of Standards
  • promote the use of OSAC-endorsed standards by the forensic community, accreditation and certification bodies, and by the legal system
  • provide insight on each forensic science discipline’s research and measurement standard needs
  • enlist stakeholder involvement from a broad community; and to [sic]
  • establish and maintain working relationships with other similar organizations. 1/
OSAC is not a regulatory agency, and it does not review the work of forensic science laboratories or practitioners. Yet one court thought otherwise. In United States v. Lang, 2/ the defendant moved to exclude testimony from "the Government's firearms and toolmark expert—M.L. Cooper." The government wanted "Mr. Cooper ... to testify about what latent prints and DNA are and the factors that influence whether a latent print or DNA are left on a surface." Evidently, the point of the testimony was to explain the absence of fingerprint and DNA evidence in the case because of "the difficulty in obtaining latent prints and DNA from ammunition and the low rate at which such types of trace evidence are recovered."

Senior Judge Raymond L. Finch of the District Court of the Virgin Islands (formerly the court's Chief Judge) denied the motion. His unreported opinion on his pretrial ruling states that
On direct examination, Mr. Cooper testified that he reviewed a Crime Scene Evidence Report prepared by VIPD Officer Don Peter, photographs of the five buckshot shotgun shells recovered from Defendant's apartment, and photographs of the black garbage bag that contained the shells. He also testified that he confirmed his resulting opinion with OSAC (emphasis added).
This was not all that Mr. Cooper, and hence the court, had to say about OSAC. Judge Finch wrote that "Mr. Cooper is currently a member of the Association of Firearms and Toolmark Examiners—an organization that according to Mr. Cooper deals with the recovery of trace evidence—and maintains contact with the Organization of Scientific Area Committees ('OSAC')" (emphasis added). The nature of Mr. Cooper's "contact" with OSAC--or is it AFTE's "contact"--was left unstated, but the court thought it was important enough to add the footnote that "OSAC coordinates the 'development of standards and guidelines for the forensic science community to improve quality and consistency of work in the forensic science community.'"

Relying on an OSAC-approved standard for an opinion is one thing. But there are no standards on how to ascertain from a "review of the Crime Scene Evidence Report and photographs" that police would not be expected to recover a latent fingerprint or an adequate quantity of DNA to analyze, and being in "contact with" OSAC does not tell the court anything about the expertise of a witness or an organization.

Other parts of the opinion are more comprehensible (but not necessarily correct). The court wrote that the testimony "satisfies the reliability test under Daubert" just because it "will be based upon his more than thirty years of experience working in the field of forensic science." Experience can be a good thing, but it is not the scientific validity discussed in Daubert. More plausibly, the court moved away from its equation of experience to science. It stated that Mr. Cooper's methodology did not have to satisfy Daubert after all. "Rather than [applying] a methodology that satisfies the requirements of Daubert," Mr. Cooper would do little more than recount his personal experience with "the recovery of trace evidence" That much, the court insisted, "is entirely appropriate under Rule 702. See United States v. McNeil, 2010 U.S. Dist. LEXIS 290, at *8, 2010 WL 834667 (M.D. Pa. Jan. 5, 2010) ('To the extent that the expert has knowledge of the frequency of firearms without latent prints, the expert [can] testify to that knowledge.')."

NOTES
  1. OSAC Charter and Bylaws, Sept. 26, 2019, Version 1.6.
  2. Crim. Action No. 2015-0013, 2016 WL 1734087 (D. V. I., Apr. 28, 2016).