‘Familial searching,’ its promise and perils

It was an unfinished slice of pizza that led to the identification of Lonnie David Franklin Jr. as the prime suspect in the Grim Sleeper murder investigation. But the pizza was just the final clue leading to his arrest.

The key break in the investigation, intermittently conducted over 25 years, came when investigators found a close — but not perfect — match between the DNA recovered at multiple crime scenes and a man being held in a California prison. Such a near-match strongly indicated that the person wanted by police was a close relative of the man in prison, and police soon focused on the man’s father, Lonnie Franklin. They put him under surveillance, obtained his discarded pizza and found that his DNA matched that recovered at a Grim Sleeper crime scene.

Four years ago, in the journal Science, we described how a data-mining technique known as “familial searching” could be used for efficient identification of possible crime suspects when traditional investigative efforts fail. The paper (which we wrote along with UC Berkeley mathematician Charles Brenner) explained how crime laboratories might benefit from searching not just for perfect matches, but also for close ones, when trying to connect DNA from unsolved crimes to the DNA of known offenders whose genetic profiles are held in local, state and national databases. Because relatives share common DNA profiles, close matches can implicate family members as possible crime suspects.

The importance of this technique was clearly demonstrated this week in Los Angeles. Yet currently in the U.S., familial searching is allowed only in California and parts of Colorado. As experience with familial searching increases, other states will probably embrace the technique. And as they do, it is imperative that policies be carefully crafted to ensure both efficiency and accuracy in case selection, statistical thresholds and follow-up testing and investigation.

Familial searching extends the size and reach of the nation’s DNA databases to effectively include the parents, children and siblings of the 8 million offenders and arrestees whose DNA profiles are already stored in databases. Additional technologies, including Y-chromosome genotyping and examination of mitochondrial DNA, can provide analyses even further out on the family tree. Extending the reach of databases to possibly tens of millions of additional individuals brings great opportunities for solving crimes. But it also raises concerns.

By utilizing these techniques, officials have the ability to reach far beyond the pool of those mandated to provide DNA samples. This sparks legitimate privacy considerations. It also magnifies concerns that African Americans and Latinos are disproportionately represented in offender databases, although this also can mean that the benefits of familial searching will accrue to these overrepresented groups. This case demonstrates that point, as Franklin and all of the Grim Sleeper’s known victims are African American.

Because of the chance that someone unconnected to a crime might appear to be related to the perpetrator, extra laboratory testing steps are always needed to narrow the list of potential suspects to avoid intrusions and conserve investigative time and resources.

But California has demonstrated how these concerns can be addressed in a way that limits the potential for intruding in the privacy of uninvolved parties, yet allows investigators to utilize an important crime-solving technique. The state limits familial searching to high-priority cases when other investigative methods have failed, requires additional Y-chromosome typing and makes use, if available, of non-forensic information in order to identify additional evidence bearing on relatedness.

The Grim Sleeper case will undoubtedly become the poster child case for proponents of familial searching around the country, and indeed, those who oppose any use of familial searching must justify not using these methods when there is lingering ongoing danger to the public. Still, states should put safeguards in place as California has done before embracing the technique.

The collection of DNA by law enforcement officials has expanded over the last decade. Initially, databases included only the DNA of those convicted of a narrow array of violent crimes. Then it was expanded to all convicted felons, and then, in California and many other states, to those arrested — but not necessarily convicted — of qualifying crimes.

Familial searching is a quantum leap because it expands the potential for DNA scrutiny to millions who have not even been suspected of a crime. And it demands a question: If it is just the capricious hand of fate that separates those of us under surveillance from those who are not, what is the justification for not creating a universal database as the only equitable solution?

Although there are many inside and outside the United States who have made cogent arguments supporting creation of a universal database, we would not support such an extension of state authority.

These are issues our system will grapple with over time. But in the interim, it is essential that states tread deliberately and carefully as they expand DNA analysis to include familial searches. Only in doing so can our public institutions both protect our individual rights while at the same time bring to justice dangerous criminals.

David Lazer is an associate professor of political science and computer science at Northeastern University and the editor of the book, “DNA and the Criminal Justice System.” Frederick Bieber is a medical geneticist at Brigham and Women’s Hospital and an associate professor of pathology at Harvard Medical School.