Anatomy of a Serial Killer

In addition to testing such systems, the Center for Human Identification is collaborating with other institutions in the effort to improve identification. It is working with the University of Tennessee, for example, to automate DNA analysis and speed up identifications for all the investigators and families tortured by a cold case. Right now, the center’s tests produce a chart of several hundred peaks and valleys that a trained forensic analyst must read one nucleotide “letter” at a time. A second analyst then reads it again to verify its accuracy. Although complete automation of the process remains a distant dream, Tennessee scientists have designed a software program that can read “perfect” sequences, or unambiguous graphics. Soon it may be able to replace the second read and thus slash personnel costs and turnaround time.

But extracting and reading DNA from unidentified remains is only half the challenge. That DNA must get linked to the right missing person. What the country has sorely lacked, Morgan says, is a central repository for information such as photos, fingerprints, dental records, DNA sequences and other identifying information on both missing persons and unidentified victims. Make that database searchable, and it becomes a profitable tool for homicide detectives. Open it to the public, and it becomes a merciful resource for the thousands who currently spend their nights combing disturbing Web sites.

In 2005 the U.S. Attorney General’s office formed a Missing Persons Task Force to develop the National Missing and Unidentified Persons System, or NamUs (identifyus.org). In 2007 the first part of the system—a searchable database of unidentified human remains—went live. Last year, the program opened up a national database of missing-person reports. And later this year, NamUs plans to connect the two, with a cross-searchable database that automatically matches the missing and the dead.

The Match

Before the NamUs database is complete, though, researchers at Fort Worth’s Center for Human Identification have to rely on meticulous information-gathering and luck. The center has put together a DNA-collection kit for family members of the missing, which it sends out free of charge to the nation’s police and sheriff’s departments. Law-enforcement officers mail cheek swabs collected from the family back to the center, where workers analyze them in batches of up to 80 to yield both nuclear- and mitochondrial-DNA profiles of parents and siblings.

As each family member’s DNA fingerprint comes off the line, it too goes through the databases to search for approximate matches among the dead. The process is spellbinding, claims forensic analyst Melody Josserand. Any of thousands of mysteries could be solved at that moment. “Even though I do searches 30 or 40 times a week, I’ve never walked away,” she says. “I sit here with bated breath.”

Josserand remembers the day in March 2006 when Unidentified Person F2775.1EC flashed across her screen. She had just uploaded family-reference sample F3352.1US, submitted by the King County Sheriff’s office. Like the reels of a slot machine, twin columns of numbers rolled down her monitor. The rows for six out of six mitochondrial-DNA base pairs flashed green. A perfect match. But mtDNA alone, she knew, wasn’t definitive. Fortunately, back in 2004, Sansom was able to pull seven markers for nuclear DNA from the victim’s bone sample. Josserand compared the family-reference sample with that. All of them matched.

Josserand retrieved the folder for Unidentified Person F2775.1EC and checked it against the file for the family-reference sample. “The metadata all matched,” she says of Debbie Deer Creek’s physical descriptors: female; approximate age, 17; weight, 125; height, 5'7". Estimated date and place of death: 8/19/1984, Missoula, Montana.