Sometimes you learn more from what you don’t understand than from things you do. We learn about what Mendel had to say about Genetics in a Zoology Lab where the fruit flies eye colors turn out just like the instructor thought they would – dominants and recessives. Then we get to Chemistry and Watson and Crick’s double helix is added to the mix, with the triplet codes carried by the nucleotides. Later in Cell Biology there’s DNA and messenger-RNA and transfer-RNA and Ribosomes. Then we read about twin studies and find out that diseases like Schizophrenia and Manic-Depressive Illness are statistically genetic, but nothing like those fruit flies in Zoology or the peas in Botany. The more we learn, the more confusing things get. Then they go and map the human genome and people start taking about SNPs, microarrays, and heat stamps. For most of us, our eyes begin to glaze over. And although we live in the first generation of people beginning to use these things to actually approach medical problems beyond the realm of "Mendelian," we might feel our eyes begin to cross and a throbbing on our forehead when we read their papers [at least that’s what I feel]. The combination of the concepts, methodologies, analyses, statistics, etc. challenges the gray matter of the initiated and uninitiated alike.
This weekend, 60 Minutes had a piece on Anil Potti MD, the Duke scientist who made a huge breakthrough in "personalized medicine" using genetic typing to select chemotherapeutic agents for specific cancers, and was in the middle of a clinical trial using his methodology. Two Biostatisticians from M.D. Anderson in Houston [Keith Baggerly and Kevin Coombs] questioned Potti’s findings repeatedly, even succeeding in getting the trial stopped for a time, but a panel from Duke investigated and Potti was exonerated. It was only when someone thought to look at Potti’s résumé claim of being a Rhodes Scholar and found out that it was a fabrication that they finally investigated deeply enough to discover that Potti’s cancer data was fraudulent. If you watched the Whitney Houston coverage and the Grammies insteadof 60 Minutes, Ed Silverman has it posted on Pharmalot:
Note that his mentor and co-investigator, Joseph Nevins, never detected the fraud, but said after the fact that the evidence was “abundantly clear”. How did something this “abundantly clear” get this far into clinical trials [100+ cancer patients enrolled] before being detected? Even by his co-investigator and mentor? Before jumping into trying to answer that, take a look at this video. It’s about 30 minutes long but well worth your time. It’s a lecture by Keith Baggerly, one of the Biostatisticians that busted Potti – about how they figured it out. Be warned, Dr. Baggerly is a fast-talking nerdish PhD who thinks you understand what he’s saying [which you likely won’t totally get], but watch at least some of it to get the flavor of the genre:
And what, pray tell, is "High-Throughput Biology"?
High Throughput Cell Biology is the use of automation equipment with classical cell biology techniques to address biological questions are otherwise unattainable using conventional methods. It exploits optics, chemistry, biology and image analysis to permit rapid, highly parallel research into how cells function, interact with each other and how pathogens exploit them in disease. It has also been called "Omics research" – the interface between large scale biology [Genome, proteome, transcriptome), technology and researchers. High throughput cell biology has a definite focus on the cell, and methods accessing the cell such as imaging, gene expression microarrays, genome wide screening. The basic idea is to take bench methods and do simply a very large number of them without losing their quality.
He’s proposing something beyond transparency – actually posting all data and all analytics [code] so it can [and will] be checked by someone off-site. He shows examples of errors [and fudging] and how difficult this stuff is to check because it is so complex. Sounds good to me! ergo Reproducible Research in High-Throughput Biology.
So now imagine this scenario. Psychiatrists are kind of new to "bench" science, and there aren’t very many who are good at it. What we have is a whole class of "armchair" scientists who think up simplistic studies and farm them out to CROs, send the data to statisticians, and coordinate getting papers written from the results often with writer-helpers. They aren’t like Dr. Baggerly – most of them. They’re arms length scientists who might wear white coats for show, but rarely are in situations to be splattered much. What would happen if they got into the world of High-Throughput Biology? Add to that, by the way, a high dose of ethically challenged. Add to that, looking for snappy results to present at the ACNP. Add to that, pharma-friendly. It’s a recipe for error, fraud, commercialized pseudoscience – chaos at best. Then consider that our journals are beginning to be populated with genome talk and genomic studies.
… genetic testing will also come into play. As you probably know, commercial companies already are marketing DNA testing. They provide a "readout" of your genotypes for all of the known coded human genes along with associations with specific diseases in the different organ systems that these correspond to, to the best level of evidence that currently exists. … there is no reason psychiatry cannot begin to use these as other fields of medicine have done. Because all mental disorders will almost certainly prove to be polygenic or multigenic, we will need a gene profile to utilize in terms of diagnostic information…
Or Charlie Nemeroff:
The findings promise to enable clinicians to match treatments to genetic subtypes. “We hope in the near future to be where infectious disease and oncology are in [terms of] individualized treatment,” Nemeroff said. “We should be able to see patients, study them in terms of their genomics, characterizing them biologically, then match them to a particular treatment regimen. “That’s what we do in oncology, and there is no reason why we can’t do it in psychiatry,” he said. “Ten or 20 years from now, we will be sending our patients to the laboratory to characterize them in terms of genetic polymorphisms and/or to an imaging laboratory. Then based on those findings, and on the clinical presentation of the patient, we will be able to do what we can’t do right now, which is to answer the question—of all the treatments that are effective for depression, what is the best one for this particular patient?”
And there are studies on the horizon:
"This application brings together researchers with extensive experience in conducting large clinical trials and experts at the forefront of the neurobiology of depression, including: clinical trials [Trivedi, Fava, Schatzberg, Nierenberg, Shelton, Gaynes, Hollon], genetics [Smoller, Binder, McMahan, Perils], neuroimaging [Phillips, Sheline, Etkin, Pizzagalli, Buckner], qEEG [losifescu, Ellenbogen], neurotrophins/cytokines [Duman, Sanacora, Turck, Shelton], clinical predictors [Shelton, Hollon, Trivedi, Fava, Nierenberg, Goodman, Yehuda], neuroendocrine markers [Holsboer, Schatzberg, Shelton, Yehuda], epigenetics [Nestler, Yehuda],and cognitive behavior therapy [Hollon, Manber, Arnow]. This team will also be guided by internationally known biomarker scientists [Holsboer, Schatzberg, Krystal, Charney, Goodman], as well as a highly qualified group of biostatisticians [Kraemer, Wisniewski, Schoenfeld]."
"by Leanne M Williams …, A John Rush, Stephen H Koslow, Stephen R Wisniewski, Nicholas J Cooper, Charles B Nemeroff, Alan F Schatzberg, and Evian Gordon"
Those are familiar names – not a very savory lot. Given their track record and the complexity of these early genomic studies, it feels like shipping a boxcar of radium to the science club at our local high school to play around with and see what they come up with. If the Genomic Scientists at Duke couldn’t detect fraud, and it took an gaggle of nerds at M.D. Anderson a number of tries to hone in on Dr. Potti, imagine the havoc when these known science clowns from psychiatry’s recent history suddenly get reincarnated as geneticists. The outcome is “abundantly clear”. Danger! Danger! Disaster looms on the horizon…
OMG! I have PTSD* already just thinking about what Perlis and Stahl will come up with.
* pretraumatic stress disorder
This is a shell game, sure, but it’s a particularly scary one because all these guys keep circling and coming back to the arena of “prevention.” Pharma loves nothing more than chronic diseases, which may or may not get better on their own or with meds–and then recur. This is the jackpot, in their terms.
So, if these guys find any genetic “anomaly” linked to any kind of mental illness, well you can just see it now, can’t you? Screening in childhood to prevent the emergence of problems that will…Or better yet–on their terms–reduce the risk of mental illness.
Like Lipitor for the brain, if you will. And you’ve seen the brouhaha over screening kids for elevated cholesterol? This is exactly the same, only much, much worse.
Amen! And one of the topics of David Healy’s book, Pharmageddon.
Here’s todays Harvard Gazette trumpeting of its latest marriage of academia and corporations: The Broad Institute’s emplacement of a psychiatrist to lead..drumroll…genetics-based psychiatric research.
YOu really have to understand the geography of this. The Broad (pronounced brode) Institute is on the MIT campus, and it is directly across the street from the Koch Center, the Whitehead Center and several other neuro/psych/genomics research centers. Whenever I walk past this area, I mean to stop and list all of the centers because I forget their names by the time I’ve walked the mile long research corridor (Google Kendall Square – Main Street on Maps, and you can take a virtual tour) Walk a 1/2 mile in the other direction, and you will be at Mass General’s door.
The Broad is the grand-daddy of them all, and the Stanley Center for Psychiatric Research has tentacles throughout MIT, Harvard and the academic medical centers (Mass General, McLean, Beth Israel Deaconness and Boston University).