FBI to Show How Genetics Led to Anthrax Researcher

A member of a decontamination team works at the Hart Senate Office Building after the 2001 anthrax attacks. (By Bill O'leary -- The Washington Post)
Buy Photo

"I don't think anybody involved knew who all the partners were," said Claire M. Fraser-Liggett, a former president of the Institute for Genomic Research in Rockville and a top scientist in her field. Her lab provided information to develop new assays, or tests, that the FBI could use to match the strain used in the attack to samples elsewhere. She said she did not know how the bureau specifically used the information.

When the deadly letters first surfaced along the East Coast after passing through the U.S. Postal Service, geneticists had not yet decoded the full genetic code, or genome, of the spore-forming anthrax bacteria, usually found in animal hides. Timothy Read, a DNA researcher who worked with Fraser-Liggett, had begun work on a sample from Porton Down, a defense research center in Britain. His incomplete research soon took on urgency.

Authorities had learned that they were dealing with one of 89 strains of anthrax bacteria. They identified their culprit as the Ames strain, cultivated in Ames, Iowa, from a sample taken from a dead cow in Texas. They had learned this by performing autopsies on the victims, which was, in itself, controversial.

The victims in Florida, Washington, New York and Connecticut died from inhalation anthrax, the most serious form of anthrax disease, and performing autopsies in such cases is highly discouraged. Once an incision is made, bacterial spores can escape from a corpse and become airborne, threatening medical personnel with infection. Military doctors are warned that in the event of a biological assault on the battleground, they should leave corpses behind.

After much debate and precautions, doctors at the Centers for Disease Control and Prevention in Atlanta finally autopsied the first victim, Robert Stevens, a photojournalist from Florida. They sent samples to the Rockville genomics laboratory.

Sequencing samples from the other victims confirmed that the Ames strain was involved in all of them, narrowing the field of possible laboratories from which the material could have come. The U.S. Army Medical Research Institute for Infectious Diseases, the Fort Detrick research center, was on a short list of possibilities.

By 2003, researcher Read, who did not return phone calls seeking comment, had focused his genetic analysis on the Ames strain. He published a paper in the journal Nature that concluded that the Ames strain DNA included more than 5 million chemical "bases" or "letters" that could serve as identifiers. He further noted that a difference of just 11 letters separated two samples of Ames that he analyzed, one a sample from the bacteria used in the postal attacks.

Other projects done at Northern Arizona University at Flagstaff and Integrated Genomics in Chicago enhanced Read's findings. Several experts yesterday said Paul Keim of Northern Arizona probably played a crucial role in the investigation. In 2004, a team he led published a paper describing a three-step method of analyzing samples that allowed them to winnow 1,067 B. anthracis samples down to 476 genetic sub-subtypes that could be distinguished from one another. Keim did not return calls seeking comment, and a university spokesman referred queries to the FBI.

The bureau did other tests to narrow the field of possible sources to USAMRIID, then concentrated its energies on scientists working within the bacteriology division. They made unannounced visits to gather samples and equipment, leading to a heightened anxiety among the division workers.

The most time-consuming process came as the FBI, in a frenzy of genetic analysis, shipped to the outside laboratories thousands of Ames strain samples from around the world. The bureau thought that it had to be able to convince a jury that its analysis of the material was foolproof.

As the investigation continued, the time and cost of doing genetic analysis plummeted, allowing the flood of samples to be turned around more quickly.

J. Craig Venter, former head of the Institute for Genomic Research, said such investigation was impossible before the recent advances. "This is just applying that same technology to forensic purposes. It's more the use of it to solve a particular criminal problem rather than [to] make advances in science," he said.

As the investigation gained focus, the samples from USAMRIID scientists offered the first real chance of pinning the material in the lethal mail to a specific scientist or team. Scientists had concluded that the formulation in the letters contained a genetic anomaly -- a flipped DNA sequence -- indicating that it was made from a combination of materials. This offered a scientific fingerprint that allowed them to compare it to formulations prepared by individual scientists.

"Just as we as humans have slight genetic variations, if there were particular anomalies used by a particular researcher, then you might be able to produce a perfect match," said a law enforcement source involved in the case.

Sources have said that the genetic trail eventually led agents to Ivins, who prepared formulations for anthrax vaccine tests at USAMRIID and other Army labs. But as elaborate and painstaking as it was, the science leaves open the possibility that someone else had access to a flask of bacteria Ivins prepared.

Sources have said that as many as 10 people worked with Ivins and could have handled his material.

Staff writers David Brown, Michael Rosenwald and Paul Kane and staff researcher Julie Tate contributed to this report.