Did FBI Scientists Identify a Single, Unique Flask?


We are slowly getting more information about the FBI’s lack of a case against Bruce Ivins. When we first learned of his suicide, we were told that he was the anthrax killer. Period. Then we got a statement that there was some technology developed by the FBI for the purpose of this case. Is this why nearly 7 years passed before an arrest was made?

The full briefing was still skimpy on the details, and not very convincing. But what we were told is that the anthrax was traced to a single flask that was in Bruce Ivins' possession at Ft. Detrick. Then, Science Magazine reported on a "close reading" of the four relevant paragraphs in the 25-page written brief. We now had a better idea of what exactly this new technology was, although the FBI refused press access to its staff scientists.

The surviving anthrax victims were recently briefed with four hours of "circumstantial," but "compelling," evidence. Now the FBI scientists have given a briefing of their own, and they have given more details about the scientific evidence. We still haven't heard anything convincing about Ivins' guilt, but the case is closed. So I guess that's it.

(As an aside, we learn in this scientific briefing that the mailed anthrax was not processed with silicon, as originally reported by the FBI.)

Two things strike me about this forensic evidence. Rather than make you read this whole column, I'll skip to the good part – these two things combine to make it clear that the "single flask" evidence was an exaggeration, to put it kindly.

First, some background on the new technology. The Ames strain of anthrax is a naturally-occurring strain of Anthrax that is being used in US biodefense research. In order to identify a source of the mailed anthrax and differentiate it from Ames strain anthrax that is found in the wild or at any of the many labs that also conduct research on Ames strain anthrax, sophisticated forensic techniques were needed.

Very similar techniques have been used for things such as The Human Genome Project. Indeed, the contractor used for the forensics is The Institute for Genomic Research – a leader in the field of genome sequencing. Ultimately, as Science Magazine pointed out, the main innovation was combining the technologies to solve a crime. Science Magazine also raised some technical questions about the data since they only had four paragraphs of explanation to work with. I won't re-iterate those questions here, and focus instead on the question of whether the data, as they are reported, actually point to a single, unique flask.

The scientists compared the mailed anthrax to samples of anthrax that are used in research, specifically, those identified as being the Ames strain. Each sample of anthrax could be classified as biologically identical or biologically non-identical to the mailed anthrax based on a sort of DNA fingerprint. This is because the anthrax bacterium replicates asexually: a single cell duplicates its DNA and splits into two new cells which have identical DNA sequences to each other. Thus, they are biologically identical.

We found out in the original written brief that the scientists tested more than 1000 samples from labs around the world. This is presumably an exhaustive sampling of all known samples of Ames strain anthrax.

Problem #1: They could necessarily only test samples of Ames strain anthrax that they already knew existed. What if there are samples of Ames strain anthrax that the scientists did not test because they did not know they existed? Aside from Saddam's non-existent WMD anthrax research, I could easily imagine any number of other governments (including our own) with highly secretive bioweapons programs hiding this information from the FBI's anthrax investigators.

All the FBI can say is that the flask in Ivins’ possession is one possible source. They can’t prove that it is the only source.

Problem #2: To allay these concerns, the FBI can claim that it is likely that it is the source since, out of 1000 samples tested, all of the ones that matched could be traced to Ivins’ flask.

What?! More than one sample matched?! Yes – a total of 8 samples matched. "B-but, b-but, they could all be traced to Ivins," stutters the FBI. We already know that Ivins was not the only person with access to "his" flask. But now we learn that, in addition to the 100 other people who had access to Ivins’ flask, there are at least 7 other people (and probably many, many more) who had access to anthrax that is biologically identical to the mailed anthrax. These 7 other samples were not located at AMRIID.

So, although we were initially told by the FBIparroting media that sophisticated scientific techniques could uniquely identify Ivins’ flask as the smoking gun, we learn that even this flimsy piece of evidence isn’t true: there are 8 known smoking guns and a theoretically infinite number of unknown smoking guns.

August 21, 2008