Rewriting the Science on DNA Mass Identification
for National Geographic News
|May 5, 2003|
After the September 11, 2001, terrorist attacks, Howard Cash attended the funeral of a New York firefighter who died in the World Trade Center. The firefighter's remains had not been recovered. As he watched a flag being folded over three helmets the firefighter once wore, Cash wondered: "How many funerals will be held without a coffin?"
Like many others, Cash felt an urgent need to help in the wake of the horrific events. Unlike many others, he was actually in a position to do something about it. As the president of Gene Codes, a Michigan-based bio-informatics firm, Cash had been tasked with inventing a new computer software program to identify the 9/11 victims based on DNA.
Over the next few months, Gene Codes doubled its staff as computer programmers worked around the clock to develop the new software. On December 13, 2001, they delivered M-FISys, short for "mass fatality identification system." The New York medical examiner made 55 positive identifications from DNA matches that day. Since then, Cash has traveled to New York almost every Friday with upgrades to the software.
The program has been remarkably successful in its somber pursuit to identify the victims' remains, many of which deteriorated so badly from exposure to fire, heat, and water that only a partial piece of DNA, if any, could be isolated. Of the 2,792 victims, forensic experts have identified 1,489 people. While the search for remains has stopped, identifications are still continuing.
Experts maintain M-FISys has rewritten the science of DNA mass identification. Robert Shaler, director of the Department of Forensic Biology in New York, which performs the lab work, called the software "a prototype" for matching DNA profiles in mass disasters. "DNA is the sexy part, but the real underpinning of [the identification process] is computer software," said Shaler.
Processing the Data
Before 9/11, Gene Codes' main business was a DNA sequencing software called "Sequencher" used by research universities and drug and biotech companies, not as a forensics tool in a disaster. The software helped scientists to work on the Human Genome Project, which was recently finished.
When Cash was invited to New York shortly after the WTC attacks, he found the DNA identification and collection system in complete disarray. Information was stored in 22 different databases, from FileMaker Pro to Oracle. "It was a horrible data processing problem," said Cash.
Gene Codes was hired on October 8, 2001, to come up with a new, sophisticated database system that could manage inventory and identify the victims' remains. Programmers incorporated some parts of "Sequencher" into the new software, but essentially had to start from scratch.
What they invented was a pattern-matching software that canin just a few secondssimultaneously sort information from three different DNA tests on human remains and compare them to DNA samples taken from victims' kin and genetic material from personal effects. Crunching DNA data that previously took two weeks now takes mere minutes.
"No one had tried to do it on a scale like this before," said Cash. "If we had written it a year before 9/11, people would have said, 'That's cool, but why would you ever need it?' The scale of the disaster changed the requirements."
Molecular Search and Rescue
After 9/11, New York City officials promised to return as many human remains as possible to the families of the victims. But the collapse of the towers burned and pulverized most bodies so badly that skeletal analysis, dental x-rays, and comparison with existing medical records were impossible.
"In other mass disasters, like earthquakes, whole bodies are recovered," said Cash. "In this case, some victims perished without a trace."
Of the 20,000 pieces of bone and tissue that were collected from the World Trade Center site, 7,066 remains have been identified5,781 by DNA only. Two of the terrorists are among those who have been identified.
While 25 percent of the remains were damaged so badly they had no DNA profile, another 25 percent only had partial identification. Forensic experts have been able to use the computer programs to combine partial fragments of DNA from related samples to create virtual genetic profiles.
Victims' families have contributed more than 40,000 genetic samples from personal effects such as toothbrushes and combs. The samples were sent to the New York State Police, while the New York City chief medical examiner received the human remains retrieved from the World Trade Center site. Upon analysis, all the information is entered into the identification database.
Some amazing matches have been made. In one case, DNA from a personal effect matched up with 208 separate remains. In some cases, identifications are gruesomely complicated because one person's bones may be embedded in another person's muscle tissue.
But there is no room for mistakes. Programmers work in pairs, sitting side-by-side in an approach called "extreme programming," constantly testing and proofreading code as it's written. One error is unacceptable.
Every week, the software helps the New York City Medical Examiner decide who goes in whose grave, said Cash. "The worst thing we could do is to identify someone incorrectly and have to go back to the family and tell them the funeral they had didn't count."
So far, there have been no misidentifications related to the software.
A Special Calling
The demands of the project have taken a toll on Cash's business. Almost the entire staff of Gene Codes is working on M-FISys. He has not been able to upgrade "Sequencher," the main income for the company.
Cash said he is answering to a special calling. "We all felt a little bit attacked," he said. "It's a privilege to work on this project. I just dread the time when the administrators will say, 'We've done everything we can,' and wrap up the project because there will still be victims who haven't been identified."
Although Cash hopes the software he has developed will never have to be used again, he knows that's probably wishful thinking. At least 11 countries have inquired about licensing the software for their own disaster preparedness programs. He is working with the International Commission of Missing Persons in Kosovo, which is trying to locate and identify some 20,000 people left missing from a decade of Balkan warfare in the 1990s.
To Cash, each identification feels "like a small victory." Every day for months, he looked for new data on the fireman whose funeral he went to. Finally, in May, 2002, there was a match. "You feel guilty about getting excited," said Cash. "But this is the most rewarding thing I've ever done."
|© 1996-2008 National Geographic Society. All rights reserved.|