What Your Second Cousin’s DNA May Say about You

Thanks to Colleen Fitzpatrick, a cold case that had baffled investigators for over two decades finally reached a breakthrough in 2014. She described the reaction in the Phoenix, AZ police department as a mixture of disbelief and elation. Fitzpatrick had just received the results of a DNA analysis obtained using a technique known as genetic genealogy, that narrowed the DNA profile down to just one individual out of over 2,000 suspects. Fitzpatrick, who has a PhD in nuclear physics, is a founder of Identifinders International, a company that provides genetic genealogy consulting services to solve forensic cases around the world.

Although genetic genealogy itself isn’t new, its application to forensics is a relatively recent development. The technique relies on Y-DNA testing, which follows the male family line, making it popular among genealogists around the world. Forensic applications use the DNA profile derived from evidence at a crime scene to search multiple, public DNA profiles available in genealogy databases, with the goal of identifying possible family members related to a suspect. CeCe Moore, founder of The DNA Detectives Facebook group, explains how genetic genealogy works in this video:

Forensic genealogy relies heavily on access to public genealogy databases. Consumer DNA testing services such as Ancestry.com and 23andMe are well-known among genealogy hobbyists, and they maintain their own collections of DNA sequence data. However, the site that gained fame in forensic genealogy doesn’t offer any DNA testing services at all. GEDmatch, unlike commercial DNA testing services, is a free, open-source database that relies on its users to voluntarily upload their DNA profile information obtained from commercial DNA testing.

At the time the site was launched, however, the founders of GEDmatch had no idea that their creation would be used to track down serial killers, rapists and others wanted by the law. Curtis Rogers, a retired businessman, and John Olson, a transportation engineer, developed GEDmatch as a side project during Rogers’ own quest to define his family tree. The site’s popularity grew rapidly, but Rogers was initially alarmed by the thought that law enforcement officials were searching a public database that held the DNA profiles of over a million Americans. Today, the GEDmatch web site contains a disclaimer about the potential use of the information by law enforcement agencies, mirroring legal language on consumer DNA testing web sites. In 2018, GEDmatch contained 1.5 to 2 million DNA profiles; it’s estimated that when that number approaches 3 million, it may be possible to determine the source of almost any DNA sample from a crime scene in the US (source). Not surprisingly, the ethical and privacy concerns raised by forensic genealogy continue to be a subject for intense debate.

Although forensic genealogy has entered the public spotlight in recent years, its precursor, familial DNA searching, has been in use since the turn of this century. Unlike forensic genealogy, familial DNA searching does not rely on open-source or publicly accessible DNA databases. Instead, forensic investigators obtain a DNA profile from a crime scene and compare it to those stored in DNA databases of known offenders, such as the FBI’s Combined DNA Index System (CODIS) or the UK National Criminal Intelligence DNA Database (NDNAD). The goal is not to find an exact match—which would be an extremely rare occurrence—but to identify profiles that are genetically similar to the suspect’s DNA, indicating that they could come from close relatives. Armed with this information, investigators could then begin systematic searches and interviews of the suspect’s putative family members.

Familial DNA searching was first applied to forensics in 2003, to solve several cases in the UK. The earliest of these was in Port Talbot, Wales, and it was a dramatic example of how the technique was used in a cold case—after the suspect had been dead for 12 years.

In 1973, two teenage girls left a nightclub in Swansea and decided to hitch a ride. They never made it home; their bodies were found in the woods near Port Talbot, and they had been raped and strangled. The team of detectives working on the case—the first serial killer identified in Wales and dubbed the “Saturday Night Strangler” by the public—soon became overwhelmed by its magnitude. More than 10,000 drivers who could have possibly given the girls a ride had to be interviewed. Further complicating the investigation, several events near the town had led to a large influx of strangers, who could also be potential suspects. By the middle of 1974, the case—then known as the Llandarcy murders--was closed in the absence of any strong leads and the evidence was shelved.

In 1990, a forensic biologist named Colin Dark with Forensic Science Services (FSS), a government-owned company, reviewed the case. He found that much of the evidence had been damaged by mice and the damp conditions in Port Talbot. Dark was interested in attempting DNA analysis, but the techniques available at the time were inadequate for the challenging task.

Ten years later, another FSS scientist, Jonathan Whitaker, was working with a recently developed, innovative DNA profiling technique called low copy-number (LCN) analysis. LCN uses the polymerase chain reaction (PCR) to amplify DNA from very small or damaged samples, making it possible to obtain a DNA profile from just a few cells. However, due to the method’s extreme sensitivity, it is also susceptible to errors from contamination and PCR artefacts, and it has gained its share of controversy over the years.

Whitaker realized that LCN might hold the key to finally unlocking the mystery hidden in what was left of the evidence from Port Talbot. Using LCN, his team was able to obtain a DNA profile of the perpetrator. However, after searching the UK national DNA database of known offenders, the team was unable to find a match. Despite that initial failure, the case was officially reopened.

A team headed by detective chief inspector Paul Bethell inherited the case archives. They had 35,000 persons of interest and knew their best hope of solving the case rested in obtaining DNA profiles, but it was clearly impractical to do so for the entire list. It took eight months to narrow down the list to around five hundred. Following further investigation, they obtained DNA samples from 353 men. None of those profiles matched the one obtained from the evidence.

Meanwhile, another cold case provided evidence that was to become vital in solving the Llandarcy murder mystery. In 1973, another teenage girl who had hitchhiked was brutally raped and strangled. Her body was found in Tonmawr, less than 10 miles from Port Talbot, three months before the murders of the girls in the Llandarcy case. The DNA sample showed that the same killer was responsible for all three murders.

Whitaker—working with Dark on the Llandarcy murders—had a stroke of inspiration. Although the DNA profile of the perpetrator hadn’t matched any of the profiles in the national database, he suggested that a close family member of the killer might be represented in the database, thereby providing a partial match—50%, in the case of a son or father. One name rose to the surface: Paul Kappen, a well-known car thief. The name was familiar to Whitaker. One of the five hundred people on the original Llandarcy list was named Joseph Kappen, but the investigators hadn’t obtained his DNA, because he’d died 11 years earlier. Under Whitaker’s direction, the team managed to obtain DNA samples from Joseph Kappen’s ex-wife and daughter. Together with Paul Kappen’s sample, these results gave them a match to 75% of the DNA profile from the Llandarcy murder case.

One final piece of evidence was required: a DNA sample from Joseph Kappen himself. The forensic team had sufficient evidence to exhume his body, despite his family’s objections. On a night that seemed straight out of a horror movie—complete with pouring rain, lightning and thunder—Kappen’s coffin was dug up and his remains sent to a local mortuary, where samples were taken from a femur and teeth. The DNA profile was a perfect match, providing the first triumph for Whitaker’s technique of “familial DNA searching”.

Across the Atlantic, initial adoption of familial DNA searching was slower than in the UK. Although the Denver, Colorado Police Department began studying the technique in 1999, California became the first state to make familial DNA searching legal in 2008. Cases in which the technique was employed had to meet two criteria: 1) the crime scene DNA profile had to be a single-source profile; and 2) all other investigative leads must have been exhausted.

The first—and arguably the most infamous—test of familial DNA searching in California was in pursuit of a serial killer known as the Grim Sleeper. The nickname derives from the 13-year break that the perpetrator apparently took after a killing spree that started in 1985.

Beginning in 2001, as part of an effort to implement new DNA analysis methods, the Los Angeles Police Department re-examined several cold cases. Continuing efforts on the Grim Sleeper case identified a DNA match among three victims killed in 1987, 2002 and 2003. Although familial DNA searching was used in 2008, it failed to provide any matches. However, the following year, the arrest of Christopher Franklin in a felony weapons charge provided a DNA sample that resulted a partial match to the Grim Sleeper’s DNA profile. Further investigation led to the arrest of Lonnie David Franklin, Jr—Christopher’s father. Lonnie Franklin’s trial began on February 16, 2016, and he was convicted on all counts of murder and attempted murder, receiving a death sentence on June 6 that year. The case captured so much public interest that a documentary film was released in 2014, Tales of the Grim Sleeper.

For more information on the Grim Sleeper case, see our interview with Rock Harmon, retired Senior Deputy District Attorney for Alameda County, California.

Forensic genealogy, which enables investigators to bypass the constraints of only searching DNA databases of known offenders, also has its share of high-profile cases. Foremost among these was Fitzpatrick’s Canal Killer case. The murders of two young women occurred in 1992 and 1993; in each case, the woman had gone for a bike ride along the Arizona canal in Phoenix and never returned. The first woman’s decapitated body was discovered in a parking lot and, later, her head was found in the canal. The second victim’s body was also found floating in the canal.

“I happened to be at ISHI in 2014,” Fitzpatrick says, “and I had the opportunity to speak to the cold case investigators there…telling them about forensic genealogy.” The Phoenix police department told Fitzpatrick they had a cold case that she might be able to help them with. They later sent her their Y-STR DNA testing results. Fitzpatrick searched several public genealogy databases to compare the profile from the case.

“I applied the software to find a match in those hundreds of thousands of profiles, and it came up with three exact matches to the name ‘Miller’,” Fitzpatrick recalls. “They had a list of two thousand suspects, and having the name narrowed it down to five.” Further investigation narrowed the search down to one individual, based on information from a forensic professional society and the FBI. The Phoenix police obtained a DNA sample from the suspect, the profile matched the one Fitzpatrick analyzed, and the rest is history.

In the years since Fitzpatrick solved the Canal Killer case, forensic genealogy has advanced to the point where even second or third cousins can provide an accurate DNA link to the identity of a suspect. In 2018 alone, over a dozen cold and active cases were solved using forensic genealogy in the US [Erlich et al., Science 362, 690–694 (2018), see Table 1]

More than half these cases depended upon the assistance of a single company: Parabon NanoLabs, where CeCe Moore is currently a Chief Genetic Genealogist. Although Moore has achieved something of celebrity status in the forensics community—with several appearances on popular television shows—Parabon wasn’t involved in the most famous case forensic genealogy case yet: the capture of the serial killer and rapist known as the Golden State Killer.

That case was cracked by a collaborative effort between forensic and genealogy teams, but it probably wouldn’t have seen the light of day if an investigator named Paul Holes hadn’t pulled out a bunch of folders from a forgotten file cabinet. The forensic genealogy analysis, conducted using GEDmatch, led to the arrest in 2018 of Joseph James DeAngelo, a US Navy veteran and former police officer.

Read more about the use of genetic genealogy in the Golden State Killer case in our interview with Paul Holes.

The adoption of forensic genealogy in the US continues to proceed at a cautious pace. At present, less than a dozen states have passed legislation permitting the use of public genetic databases by law enforcement agencies. Maryland has already banned familial DNA searches, and it may become the first state to expressly prohibit forensic genealogical analysis with legislation currently pending.

International efforts to legislate forensic genealogy mirror the concerns that consumers face in the US. “There has been much criticism of the fact that existing users of genetic genealogy service providers have not explicitly given permission for their DNA to be used for law enforcement purposes,” says Dennis McNevin, PhD, Professor of Forensic Genetics at the University of Technology, Sydney, Australia. “If public trust in the process is eroded, then the whole idea of forensic genealogy could collapse as the public withdraws its support.”

McNevin notes that developing appropriate international guidelines for the use of forensic genealogy will be a critical issue going forward, but the process is still in its infancy. Part of that process is acknowledging the limitations of the method. “Any link to a putative suspect is not sufficient evidence of their association with the crime in question,” he says. Nonetheless, he reiterates that forensic genealogy can provide valuable leads that should be used in combination with standard forensic DNA profiling and other corroborating evidence.

If wider adoption of forensic genealogy methods is inevitable, forensic laboratories around the world will need to decide how best to incorporate the technology into their workflow. McNevin recommends that agencies start by working with their own intelligence analysts to determine how the capability can be implemented within the broader intelligence cycle. They should take into account the experience and qualifications of their investigators and prepare to tackle the complex questions that will arise from the use of forensic genealogical methods. “Finally, education of investigators and district attorneys’ offices is also crucial,” says McNevin.

Clearly, forensic genealogy has the potential to provide breakthroughs for law enforcement agencies, especially in cold cases where conventional avenues of investigation have been exhausted. However, wider adoption of the technology must be guided by policies that address privacy concerns and builds public trust. Technological advancements--such as the development of detailed DNA profiles from limited samples using next-generation DNA sequencing—make forensic genealogy an even more powerful crime-solving tool. With that great power comes the great responsibility of addressing the public concern surrounding the ethical use of their data.