The Finders: Cracking Cold Cases with Genealogy, Forensics, and DNA – Part Two of a Two-Part Series – Society of Women Engineers


Each technological breakthrough turns law enforcement back to long-stored evidence, to learn the truth it’s held for years, even decades. Computer scientists and software engineers now build databases that collect and share fingerprints, records, bioinformatics, and facial-recognition data. Oxygen isotope testing reveals water signatures that reveal travel and life history, and microscopic and chemical analyses yield important clues. DNA analysis solved its first criminal case in 1986 and grows more sophisticated with each technological leap.

More tools are needed, but today there’s new hope for cold cases. Collaborations across disciplines — forensic science, DNA research, and now, forensic and genetic genealogy — introduce new ways of following data, revealing the links of DNA and kinship that result in new leads.

“Forensic genealogy is the study of identity and kinship as it pertains to the law,” Melinde Byrne, a board-certified genealogist, and a fellow and past president of the American Society of Genealogists, said. “Dealing with the living and the recently dead has ethical and legal implications that must be understood, predicted, and dealt with as they occur.”

As program director of Boston University’s genealogy program, Byrne created a forensic model for her courses. “I felt that identification of missing heirs, unknown dead, and finding next of kin to identify soldiers’ remains was only part of what could be done with standard genealogy,” she said. “As an outgrowth, so was identification of suspects in murders and other law enforcement-involved crime.”

When genetic genealogy began using data from highly specific autosomal DNA (atDNA) analysis, Byrne added not just a cold case to the course but also a “find the biological parents of” case. Today, her classes routinely solve them. And, for the past nine years, Byrne has been working a cold case of her own.

Bedford’s Jane Doe

In October 1971, a young hunter in the woods of Bedford, New Hampshire, poked his rifle barrel into a pile of sticks and leaves and uncovered a human skull. He ran to the small town’s nearby police station.

“They recovered her in the rain but didn’t get everything,” Byrne said. “A detective went back and found teeth and bullet casings, but no I.D. It was in all the newspapers, and initially people came forward, but they were looking for daughters and sisters, all of whom were eventually found.”

There was nothing left that could identify Jane Doe with the forensic methods available in 1971. Despite extensive coverage, the murdered woman, in her 30s, went unclaimed and the trail grew cold. “I learned of the case in 2008,” Byrne said, “and realized she’d been found only six miles from my house.”

Haunted by this, Byrne became determined to solve Jane Doe’s identity. Carefully building trust among local law enforcement, her help was ultimately welcomed by the police chief. “I assured him I wasn’t looking for the killer; I just wanted to name her, and whatever was learned from her name was up to him. From there it became a cooperative effort. But when we went to the state crime lab, we hit a roadblock.”

Melinde Byrne is a fellow and past president of the American Society of Genealogists. She has a B.A. in anthropology from Miami University; is currently program director of genealogy, Boston University; and is co-editor of the National Genealogical Society Quarterly. Here, she learns facial reconstruction from renowned forensic artist Betty Pat Gatliff.

Years went by, the chief retired, and Byrne realized the case would never be solved without further access. So she offered to help with another cold case that police badly wanted to solve. In exchange for access to her Jane Doe, Byrne offered expert analysis of law enforcement’s public outreach methods from a forensic genealogist’s point of view. They agreed, and Byrne presented to the state’s cold case unit and the New Hampshire State attorneys general.

Then, in 1985 and 2000, four victims were found in Bear Brook State Park near Allenstown, New Hampshire. One young adult woman and three little girls had been wrapped in plastic tied with electrical wire, stuffed in two oil drums, and dumped in the marshy woods. At the time, the “Bear Brook murders” were considered America’s most heinous crime where all the victims’ names were unknown.

“First, when you have the adult victim’s teeth, the forensic artist’s reconstruction should have the mouth open or in a smile, often the most recognizable part of the face,” Byrne explained. “Second, people don’t need to know where someone was found; they care if they knew this person. Try for intersections when the person was alive. Were you in high school with this woman? Did someone who looked like this in the 1980s go missing? Does this look like someone you’ve not heard from for a long time? Pose questions that connect her life with theirs.”

Byrne talked about phasing DNA and the important information that could be learned just comparing the four victims to one another. As she spoke, her listeners quietly made notes. “It was clear that something else was going on,” Byrne said. “At the time, I had no idea that Barbara Rae-Venter was concurrently solving the Lisa case, or how it would connect with Bear Brook.”

Mining the Data for “Lisa”

“This was not something I planned on doing.” Barbara Rae-Venter, Ph.D., J.D., is a genetic genealogy expert with a dual degree in psychology and biochemistry, and a Ph.D. in biology/biochemistry. She identified the Golden State Killer suspect in early 2018, using GEDmatch, a public database containing more than a million autosomal DNA (atDNA) profiles, to identify serial killer and rapist Joseph James DeAngelo.

In March 2015, Dr. Rae-Venter was a volunteer search angel with, teaching DNA classes online, when a webmail came in from Deputy Peter Headley in the San Bernardino Crimes Against Children detail of the sheriff’s office. “He asked if the technique we teach adoptees to find birth relatives could be used to identify someone who didn’t know who she was or where she came from.”

In 1986, 5-year-old “Lisa” was abandoned by her father in a California trailer park. A couple took her in, and another soon adopted her. Lisa grew up to be a healthy adult, married with three children. But she was troubled by dark memories of abuse, and contacted police, hoping to learn her real identity. In 2002, the man who abandoned Lisa and was assumed her father was in prison for murder. On a hunch, police ran his DNA in CODIS against hers, and found she was not his daughter but a kidnap victim. The man, then known as Curtis Kimball, refused to talk about who Lisa was or where she was from. Ultimately, police couldn’t further trace Lisa’s identity without the help of a genealogist who understood DNA.

Dr. Rae-Venter, who not only understands genetic science, but also the law, volunteered to help, and was given Lisa’s Ancestry DNA profile and also had her tested with 23andMe. She then uploaded each file to both GedMatch and FamilyTreeDNA. After an exhaustive comparison with public databases, and through traditional genealogy’s deep study of public records and DNA screening involving a possible 25,000 relatives, she was able to name Lisa’s mother, Denise Beaudin of Manchester, New Hampshire. She began building Lisa’s family tree.

Then the story began taking twists and turns. “When I told deputy Headley, he searched for Denise in New Hampshire’s databases and said, ‘No, she doesn’t exist.’ I said, ‘Yes, she does. She’s in her grandmother’s and brother’s obituaries, and she’s also in my tree.’

“We learned from Lisa’s New Hampshire relatives that Denise went missing from Manchester in 1981 with her boyfriend and 6-month-old daughter. Since she hadn’t been voting or driving since then, she was no longer in law enforcement databases — that’s why they didn’t find her. The obituaries I found were published long after she disappeared. It was pure serendipity that her family listed her as a survivor — hoping she was still alive.”

While building Lisa’s family tree, Dr. Rae-Venter found her maternal grandfather, asked him who Lisa’s father was, and got a name: Bob Evans.

“Bob Evans was not one of my candidates for Lisa’s paternity. So the question became: Who is he? On a hunch, detectives sent the grandfather a photo of the man who abandoned Lisa in the trailer park. He said, ‘That’s him.’ Suddenly, the circle closed — Bob Evans was Denise’s boyfriend in New Hampshire.”

Denise, Evans, and Lisa (whose birth name is Dawn Beaudin) were never reported missing. Evans had told family and friends they were going into hiding because they owed people money.

Manchester, New Hampshire, where the three were last seen together in 1981, is only 16 miles from Bear Brook State Park.

Realizing this, investigators matched the timelines and location of Evans’ employment as an electrician and the geography of the area, and named him as the prime suspect in the Bear Brook murders.

“Forensic genealogy is the study of identity and kinship as it pertains to the law. Dealing with the living and the recently dead has ethical and legal implications that must be understood, predicted, and dealt with as they occur.”
– Melinde Byrne, certified genealogist; fellow and past president, American Society of Genealogists; and program director of genealogy, Boston University

Solving The Multiple Aliases Of A Killer

Dr. Rae-Venter was soon asked by New Hampshire law enforcement if she could help identify the Bear Brook victims, the very case Melinde Byrne had presented on.

“You don’t normally have the murderer’s alias before you know who the victims are,” Dr. Rae-Venter said. “So they also asked me to identify who Bob Evans really was.

“He’d been stealing identities at least since 1981, so part of his profile was looking for someone who had disappeared from the databases. We knew he was an electrician. He had dropped other little details in prison — he’d lived in Hawaii, Texas, and California, and was born in Wyoming. But who knew if any of that was true?”

Mining the data and armed with Evans’ DNA, Dr. Rae-Venter built Evans’ family tree, relying on matches with his U.K. ancestry, finding enough data to pinpoint a possible common ancestor. “We studied how that ancestor’s descendants were showing up on birth, marriage, and voting records. Eventually, we ran a couple of names through the regular police records — people who had been moving around.”

One name hit all the marks. He had married in Hawaii, lived in Texas and California, and divorced there. “We requested a copy of his divorce papers from San Mateo County, and hit the jackpot — his occupation was electrician.”

Still, there was no Bob Evans in law enforcements’ databases. But Dr. Rae-Venter found yearbook pictures from his high school in Arizona. Investigators ran those pictures, matching him to the mug shot of a man arrested and imprisoned for another crime.

Barbara Rae-Venter, Ph.D., J.D., is a retired intellectual property attorney who specialized in the patenting of biotechnology inventions. She has a J.D. from The University of Texas at Austin Law School, a B.A. double major in psychology and biochemistry, and a Ph.D. in biology (biochemistry) from the University of California, San Diego. She was instrumental in identifying the Golden State Killer in 2018.

He was born in Colorado, not Wyoming. “Bob Evans” was Terry Peder Rasmussen, and he had died in 2010 while in prison for the California murder of Eunsoon Jun.

In 2017, thinking of the Bear Brook victims on the other side of the country, Dr. Rae-Venter asked for Rasmussen’s DNA. “We had identified three of Rasmussen’s children by another marriage, so police got a sample from his son. We ran that in the FBI’s CODIS database and it was a match.”

Was Lisa’s mother, Denise Beaudin, the adult victim from Bear Brook? CODIS was run against all four, and she was not, nor was Denise related to the girls. But the unknown young woman from the oil drums was mitochondrially related to two of the three girls found with her. Oxygen isotope testing had long ago identified the three as being from the same area, through the chemical signature of a water supply unique to their town. Traces of strontium, nitrogen, and carbon in their bones matched them to a common diet. Those tests also revealed that the third girl was from somewhere else.

Then another shocking twist: The third girl’s autosomal short tandem repeat (STR comparison in CODIS) matched Rasmussen’s. “She was his own daughter,” Dr. Rae-Venter said. “So it’s even more astonishing that he didn’t kill Lisa. Basically, the people who took her on a trial adoption in 1986 saved her life.”

A Technological Leap Answers a Decades-Old Question

With Rasmussen’s and his relationship identities solved, three mysteries remained, and the biggest challenge was retrieving usable DNA from the remaining Bear Brook victims. “We did two or three extractions from bone, and a liver extract, and what little we could get was very low quality. In the first sequencing, the amount of human DNA was only 2 percent — the rest was bacterial. Theoretically, you could amplify that, but only at extraordinary expense.”

Then, Dr. Rae-Venter remembered reading about a case where atDNA was obtained from rootless hair for the first time, using technology developed for use by paleontologists.

Forensic composite of one of four victims found in Bear Brook State Park, near Allenstown, New Hampshire, in 1985 and 2000. Their identification process continues in 2019.
Bedford, New Hampshire’s Jane Doe, found in October 1971. This reconstruction was created on the original skull by a police artist.

“Nobody had ever gotten at DNA from the hair shaft before.” Dr. Rae-Venter explained that as hair grows, the cells undergo apoptosis, or programmed cell death. As that occurs, the nuclear DNA gets chopped up. The fragments are so small, some of them are only about 45 base pairs, and even polymerase chain reaction (PCR) techniques can’t detect at DNA in the hair shaft.

“I contacted the lab and they agreed to test hair from the Bear Brook victims, which, luckily, was plentiful. After adding a bleaching step to remove the melanin that interferes with DNA extraction, 10 strands of the woman’s hair was sent to the lab. Because the nuclear DNA fragments were so tiny, whole genome sequencing was performed and an algorithm applied, extracting enough data to create a file uploadable to GedMatch.”

Thirty-three years after the crime, the investigation continues.

Mysteries That Remain

Bear Brook still ripples outward, and the multiple aliases of a transcontinental drifter mean there may be more “Does” scattered in Rasmussen’s wake. Dr. Rae-Venter, now publicly credited with identifying both Rasmussen and the Golden State suspect, has been inundated with calls for help. Her backlog of 70 cases includes many involving crimes against children, for whom she particularly craves justice.

For Byrne, the methodical search for her Jane Doe’s identity continues. Fifty years later, the wooded area where she was found is now a freeway ramp. Currently under nondisclosure, Byrne cannot discuss recent progress in the case, but says, “I believe there is more than enough data to solve it. State-of-the-art crime investigation in 1971 didn’t have the technology and collaborative approach that it does today. Now, every cold case can be solved — if you have DNA.”

For deeper background on Bear Brook, listen to this excellent podcast series from New Hampshire Public Radio.

Read part one of this series — our interview with genetic genealogy expert Colleen Fitzpatrick, Ph.D., also in this issue. Dr. Fitzpatrick is a fellow of the Society of Photoinstrumentation Engineers and an associate member of the American Academy of Forensic Sciences. In this story, she discusses the DNA data that link us all and how it can be shared, while touching on historic cold cases involving Amelia Earhart and Abraham Lincoln.

“Extracting atDNA from the hair shaft was not possible before. The implications of this technology for solving cold cases worldwide are just huge.”

– Barbara Rae-Venter, Ph.D., J.D., retired intellectual property attorney and genetic genealogy expert

Reverse Engineering a Family Tree

Y- and mtDNA are basically unchanged from one generation to the next, and matches mean you may have a common ancestor from hundreds of years ago. An identical match might narrow that to the last two centuries, but is not definitive. Autosomal DNA (atDNA) goes through recombination from one generation to the next, and is a random 50 percent mix of the parents, and a different mix for each sibling. Its power lies in DNA segment triangulation. First, identify at least three people who match on a given DNA segment (the person of unknown parentage counts as one). If they are second cousin matches, then they share great-grandparents as their most recent common ancestor (MRCA). Since they all share that DNA segment, the person you’re trying to identify must be a descendant of those same great-grandparents. Then, build a reverse tree down from the MRCA to the time period you believe the unknown person was born. Finally, evaluate who is found in order to identify the person of interest, using physical characteristics, geography, etc., to narrow down a list for “target testing” of family members to see if you have identified the person correctly.