Advertisement
Previously known only as the "Buck Skin Girl," Marcia L. King was recently identified as the young woman found murdered in Troy, Ohio in 1981. The breakthrough in the 'Buck Skin Girl' case was made possible through a new initiative called the DNA Doe Project. (Photo: Courtesy of Colleen Fitzpatrick)

After decades of work, and hundreds of leads, it was two experts at their computers, in the middle of the night, who cracked the case.

A four-hour flurry at their computer screens gave a name and a face—and started a homicide investigation—in one of the most infamous of Ohio cold cases: that of the Buck Skin Girl, found in a roadside ditch near Troy, Ohio, in 1981.

Starting late the night of March 28, Margaret Press and Colleen Fitzpatrick took the highly-degraded incomplete DNA profile, and started combing a public genealogical database. Within a short time, they had a staggering hit: a first cousin once removed.

They found the woman on Ancestry.com. Then they found the corresponding family tree, where they could click around on the grandparents, and the great-grandparents. Then they split up and worked backward, toward descendants. 

“By a stroke of luck, I opened the name Marcia,” Press recalls. “And a death field was filled in. It said, ‘Death—Unknown. Missing—Presumed Dead.'”

By 3 a.m. on March 29, they were looking at the Buck Skin Girl’s real name: Marcia L. King.  After nearly four decades of searches, interviews, advanced forensics like palynology and stable isotope analysis, and in-depth looks by detectives and experts alike, they had found it.

“You can’t get a more-immediate confirmation in a case like that,” said Press.

“After 37 years, to solve it in four hours—I would say it was not expected,” said Fitzpatrick.

The two women are the co-founders of the DNA Doe Project. King’s identification marks the high-profile debut of the nonprofit group that started last year. A handful of other cases—similarly baffling, and unsolved—are being put under the magnifying glass using the same methods employed in the Buck Skin case. Advanced DNA analysis of degraded samples, complex sequencing and the use of public DNA databases could provide a link to the unnamed and unidentified mysteries that have lingered for years, and perhaps decades, the women told Forensic Magazine in a recent interview.

“The key is work with degraded DNA,” said Fitzpatrick. “This is not just a breakthrough on genealogy at work. We have adapted the existing tools, and we have developed new tools, and to apply the genealogy in a new context.”

(From left to right) Detective Steven Hickey, forensic genealogist Colleen Fitzpatrick and genealogist/researcher Margaret Press. (Photo: Courtesy of Margaret Press)

PULLING OUT ALL THE STOPS

The Buck Skin Girl remains were long gone, buried in a Jane Doe grave just weeks after she was found in a roadside ditch in April 1981.

But what remained in storage in an evidence room was a tube of blood taken during autopsy.

There was a catch: the blood had been kept at room temperature, and was not refrigerated in the ensuing 37 years. Several laboratories said the DNA was not viable, according to Steve Hickey, the lead detective on the case, from the Miami County Sheriff’s Department.

But Elizabeth Murray, a forensic anthropologist in Ohio, brought the tube to DNA expert Weining Tang, of AMD Biotech.

Murray, who had been assigned the case through the National Missing and Unidentified Persons system, or NamUs, had sought the pollen and isotopes work in recent years. The Buck Skin Girl case featured good AFIS fingerprints, unique dental status and other identifiers that should have brought an ID long ago, Murray said.

“I just couldn’t understand why this case couldn’t be solved,” Murray said.

But it was the chance meeting at the 2017 American Academy of Forensic Sciences conference, in which she had heard Fitzpatrick’s genealogy talk, that prompted the anthropologist to reach out immediately—even before the talk had finished.

“I said, ‘I have the cases if you have the methodology,'” said Murray. “We had all the other stuff—why not try this?

“My role in this case was as a broker,” she added.

Forensic anthropologist Elizabeth Murray worked on the Buck Skin Girl unidentified person case before connecting with Colleen Fitzpatrick of the DNA Doe Project, in the hopes of using new methods to identify the murdered woman. (Photo: Courtesy of Elizabeth Murray)

SNPs

Between 50 and 75 percent of the crucial 600,000 to 900,000 SNPs in King’s DNA had survived the decades of storage without refrigeration, said Press and Fitzpatrick. Those SNPs are the same sets used by direct-to-consumer services such as Ancestry and 23andMe, said Press.

“Fifty to 75 percent of those points were left after they massaged it to get us the data we need,” said Fitzpatrick, a forensic genealogist and scientist who is well known for her work helping to solve the infamous Phoenix “Canal Murders” several years ago. “We got enough marbles in that jar to make it work.”

“We throw out 99 percent of the genome that we got,” said Press, who is a genealogist, researcher and writer.

A facial reconstruction of the Buck Skin Girl created by the National Center for Missing and Exploited Children before she was identified as 21-year-old Marcia King. (Image: Courtesy of the National Center for Missing and Exploited Children)

TO FAMILY TREES

The data points they had were uploaded into GEDmatch, a public database used by many professional genealogists searching for family trees. GEDmatch is for those who have their data sequenced in Ancestry or other services—but who want to see if it matches with people outside their own service.

GEDmatch is smaller—its 800,000 profiles are about a tenth of the estimated 8 million in Ancestry.

However, its swath is wide enough to produce huge breakthroughs, said Press and Fitzpatrick.

Using a script developed by Greg Magoon, a senior research engineer at Aerodyne Research, they could take the incomplete profile of the Buck Skin Girl and search among the existing members in this case.

The script essentially pulls out of the degraded sample as much of the typical data found in 23andMe or other services—and then puts that into GEDmatch, said Press. From there, the “template” sample is then run typically through GEDmatch’s proprietary software.

Fourth cousins can be found through such a search. But some studies have indicated that the finding power can extend even to ninth cousins, said Fitzpatrick.

With Buck Skin Girl, it came much, much closer to home.

Within hours, they had found a match close enough to be a first cousin once removed. From there, they found a family tree that showed a Marcia King—missing and presumed deceased. And the two women were confident that they had found the true identity of the Buck Skin Girl, after decades of other leads that reached dead ends.

(Fitzpatrick and Press also credit the many hours spent by DNA Doe Project volunteers who helped narrow down this and many other searches).

Hickey, the detective, told Forensic Magazine that the case was forensically “groundbreaking,” because of the breakthrough made by Fitzpatrick and Press. The unrefrigerated blood sample yielded DNA to the two experts—and then led to a genealogical match in GEDmatch, he said.

“I encourage any law enforcement agency who had a cold case like this to do everything they can,” said Hickey.

The pollen and isotope analysis was not far off, either, he said—it showed that King traveled a lot in her last years of life. So far, authorities can point to her being in Louisville, Ky. before she died—and Pittsburgh, Pa., in March 1981, in the month before she was killed.  

Now they have an active homicide investigation, Hickey added.

“We can start a victimology now,” Hickey said, adding that tips have already been coming into the sheriff’s tip line at (937) 440-3990.

Marcia L. King, 21, was found murdered on the side of a road in Troy, Ohio in April 1981. She went unidentified for 37 years before degraded DNA from a blood sample taken at autopsy was matched to a distant relative, whose profile was uploaded to online genealogy database Ancestry. (Photo: Courtesy of the Miami County Sheriff's Office)

FUTURE CASES

The DNA Doe Project announced on its Facebook page in February that they were looking at Buck Skin Girl—but also four more. They are: the unknown identity of a man known as Joseph Newton Chandler III, who died by suicide in 2002 (an Office of the U.S. Marshal for the Northern District of Ohio case); the Belle in the Well case, an unidentified deceased woman found in 1981 in Lawrence County, Ohio; the Mill Creek Shed Man case in Snohomish Co., Wash.; and the death of a man known as Lyle Stevik, a man who hanged himself in Amanda Park, Wash., in 2001.

With such challenging samples, there is a mixed chance of success. Five active cases are under review currently. But several others were “non-starters,” especially because unidentified bones had been boiled to remove soft tissue for analysis. That had degraded the genetic material beyond recognition, even under the DNA Doe Project’s exacting methods.

The DNA Doe Project has been contacted by dozens of law enforcement agencies. Where the nonprofit heads next after its first big breakthrough is anybody’s guess—but Press and Fitzpatrick said they plan to continue the work on the toughest cases they can find.

“It has a life of its own already,” said Fitzpatrick. “There are so many missing people, and so many unidentified people. And surely some are in both categories—and nobody knows it.”

Advertisement
Advertisement