MassGenomics

This week we had a lecture from Claire Fraser-Liggett on the FBI’s investigation of the anthrax attacks of 2001, a.k.a. the “Amerithrax” investigation. The seminar was hosted by the Midwest Regional Center for Excellence, a research network for biodefense and emerging infectious diseases, headed by WashU’s Samuel L. Stanley Jr. Dr. Fraser-Liggett had three main draws, I think, that made it a full house. First, the scientific draw, because she’s one of the foremost microbiologists in the field. Second, the celebrity draw, because she was married to J. Craig and used to head TIGR. Finally, the subject-matter draw: Anthrax.

We’re only just now getting Dr. Fraser-Liggett’s insider’s view of the Amerithrax investigation because she was under a six-year nondisclosure agreement with the FBI. Evidently, she and her colleagues were brought in because of their microbiology expertise and because they happened to be sequencing strains of B. anthracis. They co-authored a major paper last year on the Global genomic population structure of Bacillus anthracis, the “Anthrax HapMap” as she called it.

Let There Be Microbial Forensics

The timeline and targets of the 2001 anthrax attacks are an interesting story in their own right and nicely chronicled on Wikipedia’s Amerithrax page. After the attacks, several samples were collected from tainted envelopes, while one (that was sent to the National Enquirer publisher) was isolated from the poor fellow that died of it (no envelope was ever found). The investigation team quickly determined that all of the samples derived from the “Ames” ancestral strain of B. anthracis, which was isolated from a dead cow in Texas over two decades ago. Intriguingly, the USAMRIID microbiologists working on the samples noticed that the attack samples had five distinct “morphotypes” when grown on plate agar. They creatively named these A through E.

It seems that the anthrax pathogen has very little genetic heterogeneity. Yet there were rare, subtle sequence variants among the morphotypes that provided molecular markers for some microbial forensics. A particular set of four mutations identified all of the attack morphotypes, and was used to screen against the FBI’s repository of all known laboratory strains of B. anthracis. Just 8 of 1100 repository samples had all four mutations, and the investigators traced their origins back to a single flask of B. anthracis, labeled RMR-1029, that had been prepared in 1997 for large-scale vaccine trials. Ladies and gentlemen, I give you the murder weapon.

Closing In on the Suspects

It’s safe to say that the FBI was taking a many-pronged approach to investigating these attacks, of which microbial forensics was just one. They had in fact identified some “persons of interest” by scientific and non-scientific means. One of the accused had his reputation basically ruined by the press, and later slapped the DOJ with a lawsuit that was settled for something like $5.7 million. But the genetic signature of the attack strains was probably the most compelling evidence, the fingerprint on the smoking gun. The FBI spotlight came to shine on a top bio-defense researcher who’d worked for the last 18 years at Fort Detrick. This individual had participated in a working group for protocols of collecting samples for the FBI’s repository. He had also submitted some samples to that repository that were very closely related to the strain used in the attack. The FBI raided his lab and confiscated a flask that contained all four mutations of the attack strains. After being informed by the FBI that he was to be charged for the anthrax attacks, the researcher committed suicide.

Unanswered Questions

Dr. Fraser-Liggett brought us to a series of unanswered questions about the attacks, the investigation, and the pathogen itself. Should we start building repositories of other dangerous pathogens now? Or, did an emphasis on bio-defense after the September 11th attacks make us less safe by putting the materials of biowarfare into many hands? Many questions less dramatic, but more scientific, surround the pathogen itself. Are there other mutations to be found, perhaps with next-generation sequencing technologies? What is the effect of purifying selection on such variants during anthrax’s passage through the host? I have no doubt that Dr. Fraser-Liggett and her colleagues are already racing to answer these questions. This time, perhaps, in a lab far from terrorism and suspicion and non-disclosure agreements.

Last Friday we had a seminar from David I. Smith, who was in town to run the St. Louis marathon. Evidently his goal is to run a marathon in every state. He gave us a seminar of next-gen sequencing applications at the Mayo Clinic, which include gene expression profiling, rearrangement/CNV detection (with paired-end), chip-seq, methylation, microbial genomics, and characterization of large genes.

Their main focus, however, is using next-gen sequencing for rare variant detection. The goal is to develop clinical assays for early detection of cancer from blood, saliva, stool, or urine samples by looking for somatic mutations carried by a small number of cancer cells. The ratio of mutant to wild type signal is presumably low at early cancer stages, but theoretically detectable with sufficient cancer coverage.

Another clinical application of this approach is the detection of “heteromorphisms” in mitochondrial DNA. Here, the idea is to detect diseases associated with mitochondrial mutations (encephalopathies, neuropathies, oxidative phosphorylation disorders, etc.). Apparently mtDNA has a high mutation rate and propensity to accumulate somatic mutations due in part to poor DNA repair mechanisms. While the population size within a single patient is large, the genome is small enough (16kb) to be covered at high depth by next-gen technologies.

Finally, Dr. Smith presented some of the work done at Mayo on a transcriptional profiling technique that uses 454 sequencing of 3′ tags. They’ve developed approaches to profile both the polyadenylated and the non-polyA transcriptome at a cost of around $3,000 per sample.

Overall, I found Dr. White to be an engaging speaker. He also seemed very knowledgeable of the current state of various next-gen platforms, and by dropping names such as Eland, Maq, and SSAHA, convinced me that (while they only have a couple of machines) they’re pretty serious about next-gen sequencing informatics at the Mayo Clinic.