New York is pursuing a missing-persons investigation greater than any ever undertaken, attempting to match random body parts from the WTC disaster to more than 2,800 victims' names. At the heart of this quest, a Berkeley mathematician writes, and rewrites, software designed to sift through mountains of genetic code. This most macabre of mysteries -- to whom did these thousands of tiny, faceless pieces belong -- is also fantastically complex. Given the intricacy of every bit of genetic code, it's something akin to finding patterns among the stars in the cosmos, then matching them, precisely, to patterns among grains of sand.
By some measures, the World Trade Center DNA project is also a gargantuan exercise in futility: Of approximately 900 DNA profiles with a potential to be matched, Berkeley forensic mathematician Charles Brenner says investigators have so far matched only a quarter of these remains using DNA analysis. Some 1,800 victim families have requested death certificates based on non-DNA evidence showing that their relatives were in the buildings at the time of the disaster.
Despite its limited success, this mass genetic-matching effort is actually far from pointless. Following massacres in Africa, South America, and Eastern Europe, where international experts spent years exhuming murder victims to help families learn what had happened to loved ones, human rights experts now believe that there's nothing like truth to help people recover from atrocity.
If the World Trade Center DNA project has an elegant goal, its execution -- digging through piles of rubble to find body parts, then knocking on doors to ask for DNA-laden mementos to use in the matching process -- has been a particularly inelegant exercise for Brenner and his colleagues.
Still, it's the kind of monumental mess of the sort Brenner has spent more than a decade preparing for. Perhaps the world's only freelance forensic mathematician and one of the planet's foremost experts in solving the numerical equations required to match DNA samples, Brenner has traveled to Argentina, where he wrote software that helped establish the parentage of some 300 infants given to childless military families after their biological parents were murdered by agents of that country's 1970s-era military dictatorship. Brenner wrote algorithms that helped swiftly identify all 229 victims of Swissair Flight 111, which crashed off Nova Scotia in 1998. And Brenner's mathematical tools helped resolve paternity questions related to Larry Hillblom, the Bay Area multimillionaire and DHL delivery service founder who sired at least a half-dozen heirs in liaisons with teenage girls in Micronesia and East Asia.
As with these previous cases, much of the WTC effort involves ordinary detective tasks such as digging up remains, categorizing evidence, and knocking on doors. But in New York, as elsewhere, matching genetic information winds up being above all a mathematical puzzle; the heavy lifting of DNA fingerprinting science lies in calculating the mathematical probability that seemingly random markings in one genetic fragment are biologically related to similar pieces of DNA code in another. Seeking to match thousands of body parts with thousands more DNA samples and/or family members presents a forensic number-crunching task like no other tried to date.
Given the magnitude of the job, ordinary DNA techniques fell far short. Matching 14,000 body parts with material from thousands of family members would take years to complete if standard methods were used, Brenner says. "You would be swamped by plausible relationships between nonrelatives," says Brenner.
At first it was far from clear which techniques would prove most effective in sorting information from the World Trade Center remains. Investigators convened a forensic science summit of DNA experts from around the world. Experts considered software used by the FBI, but it turned out the FBI software was of little use matching massive amounts of DNA information from relatives.
"Starting with the first meeting, I started agitating to get the data," says Brenner, who flies again this week to Albany for a meeting of the committee formed to advise the World Trade Center DNA project. "Not knowing any more dignified way to put it, I spoke up at the meeting and said, "Give me the data, and I'll solve the problem.'"
New York officials relented, and Brenner went to work, eventually writing a DNA-specific version of today's intelligent Internet search engines. "If you're a little more clever, and let's say you have two family members in a given reference family, two living references, and you make sure both of them appear to be related to a given victim before the computer gives any signal, then you filter out the majority of the false matches," he says.
"After I had the data, it only took me a week or so to extract the available information from it, which I passed back to New York," he adds. "And they gave me some more sample data."
Looking past Dr. Brenner's excitement at the intellectual challenge presented by the World Trade Center DNA project, it's possible to question the wisdom of spending thousands of man-hours on a forensic effort doomed to achieve only modest investigative success. Last month New York officials asked family members to scour their homes for more DNA-marked remnants of loved ones -- dirty clothes, hairbrushes, anything. But even this second call for genetic material is unlikely to produce a thorough catalog of the World Trade Center dead. Too many people were made cinders by the heat of the jet fuel explosions involved; the cataloging of remains has lacked the kind of precision that might have been possible were this a smaller crime scene; in many cases, producing sufficient genetic material from victims' families won't be practical. Out of fewer than 900 fragments with DNA profiles that provide the potential to be matched with those of family members, the New York Medical Examiner's Office has so far only produced a couple of hundred actual matches as a result of DNA, Brenner says.