genomeboy.com

Thank you, James T. Kirkus. Note that the official pub date has since been moved up to 2 November 2010.

From the Jewish Daily Forward:

“If you test positive for something and go online and start reading, you have to know what’s accurate information and what isn’t,” said Rhonda Buyers, executive director of the National Gaucher Foundation. “There’s a lot of junk online, and you could be reading inaccurate information and say, ‘Oh my God, it’s fatal. I’m going to die.’ There could be treatments that you’re not hearing about.”

But Misha Angrist, an assistant professor at the Duke Institute for Genome Sciences & Policy, said that requiring health care providers to serve as gatekeepers for genetic information is unrealistic. “Within not that many years, the technology will be such that people can do this in their garage, and to say that you should have to go through a doctor is naïve,” he said, citing the limited number of medical geneticists and genetic counselors. “The infrastructure of medical genetics is not up to holding everybody’s hands.”

Well, it’s not only the folly of the status quo that singes my shorts. It’s the cockroach-like survivability of good old-fashioned genetic exceptionalism. Notice how the Gaucher Foundation director assumes that the reaction to one’s genetic test results will be horrific, worst-case-scenario, “I’m-going-to-meet-my-maker.” She assumes that the vulnerable googler will come upon erroneous information and draw the wrong conclusion. She assumes that that person will suffer for her actions. To paraphrase all the DTC haters, “show me the data.”

***

Last week I had minor, elective surgery. (I’ll spare you the details; it was painful, but as far as I know, successful and uneventful.) Two things to note:

• Before taking acetaminophen with codeine, I checked my genome to see if I carried an allele that made me less likely to respond to codeine. The CYP2D6 literature is a bit of a mess, but I concluded that I probably metabolized codeine normally and so took some Tylenol III feeling fairly certain I was not wasting my time in doing so. And I wasn’t (at least until I got my oxycodone scrip filled…yeah baby).
• A couple days after the surgery, I got an email from the hospital: my pre-op lab results could be downloaded. “Cool!” I thought. Perhaps the end of the buggy whip was nigh!  I opened up the pdf and looked at my various blood cell counts. My white cell count was bolded and *asterisked*. Hmmm…why was that? No explanation. No mediation. No hand-holding! Clearly the next step was to jump off a building. A few minutes of web-surfing assured me that my WBC count was not that high and probably suggestive of a minor infection the week before.

But the point is this: I don’t remember consenting to this data exchange.  Thus, I was sent information online:  1) that I probably did not ask for; 2) without any guidance; and 3) for which there was no a priori reason to assume comprehension on my part. And guess what:  it was okay! I could have chosen not to open the file or I could have contacted someone at the hospital and asked them to walk me through it. Instead I reckoned I could handle it…and I did. Next.

I have to believe that these sorts of lab test results are being dispensed in this way everyday to thousands of surgical patients who are, like I am this week, laid up and not at their best. To my knowledge, however, neither the Energy and Commerce Committee, nor the FDA, nor the GAO, nor the CDC nor any other would-be guardian of the public health has ever made a peep about this practice. (I imagine they are too busy chasing purveyors of Reckless Legs Syndrome risk estimates.)

In all seriousness, I was thrilled that my hospital opted to share my data with me via the simple act of sending me a pdf. The time has come to re-evaluate our reflexive response to information about ourselves, i.e., that we must be protected from it at all costs and/or spoon-fed it via some anointed bureaucracy. Because frankly, if those are the hurdles, then many of us will simply never bother. After all, we have more pressing questions. Like WHERE’S MY PERCOCET????

I can’t help but wonder if these guys are forced to blog about the FDA as some kind of punishment. That said, they are doing a brilliant job so far. Today’s post features an extended riff on the ongoing regulatory kerfuffle surrounding consumer genomics companies. What struck me was how markedly the FDA’s stance has changed from the heady and idyllic days of 2007-2008:

Federal regulators were favorably disposed to the industry at the outset. Before launching their product, 23andMe met with then FDA Commissioner Andrew von Eschenbach, who “encouraged [them] to proceed”. Several more startups were even approached by senior FDA officer Lawrence Lesko to officially collaborate on a study of genetic influences on drug response. Lesko, the Director of FDA’s Office of Clinical Pharmacology, gave a presentation at Temple University where he mentioned 23andMe’s potential to improve “access to genetic testing” and enable a “new, educated generation of health care providers and patients.”

“If I’m to take a consumer genetic test, how do I know if I can trust the results, and should I be scared of what I might find out?”

Mary Carmichael of Newsweek asked me and a number of others to reflect on this question. And so I have.

• The short answer to the first part is, you may never be able to trust ALL of the results. Personal genomics companies type customers for several hundred thousand or a million variants (single-nucleotide polymorphisms or SNPs) and return results on a few hundred at most. My sense is that those few hundred are selected based on 1) the scientific validity of their association with particular traits; and 2) the presumptive interest of customers in those particular traits. But just because an association between a genotype and a trait is “real” doesn’t mean it’s predictive. Height, for example, is highly heritable, but it is probably mediated by thousands of genes. The other issue, as I’ve pointed out previously, is that the consumer genomics companies have thus far declined to issue standards. They use different SNPs in many cases. To assess multiple sclerosis risk at present, for example, 23andMe looks at two SNPs: one in a gene encoding a growth factor that sends signals to the immune system (IL7RA), and one that is tightly correlated with a SNP (HLA-DRB1) in the HLA complex, which is a region on chromosome six that harbors many genes involved in immune function. By contrast, Navigenics looks at those two plus another SNP in the IL2RA gene, which is important in T-cell activity and is a suspect in other autoimmune diseases. Both companies tell me that my lifetime MS risk is slightly lower than average, but the two numbers are not identical (23andMe: 0.24%; Navigenics: 0.17%). Meanwhile, a collaborative group of scientists has developed an algorithm incorporating 16 SNPs that is modestly predictive of MS risk. And SNPedia, the public wiki that collates SNP associations with human traits, lists more than 25 SNPs contributing to MS risk. Thus, the current level of confidence one can have in a SNP-based prediction of one’s chances of developing MS is weak at best.

• Should you be scared? Well, clearly in the case of MS, unless your neurologist tells you otherwise, you should not. But some consumer genomics companies return results on certain genetically mediated traits where the results are not so ambiguous. 23andMe, for example, returns results on three mutations that, if present, render women highly susceptible to breast and/or ovarian cancer. This is done on an opt-in basis: these data are not returned automatically with the rest of one’s 23andMe results. Some bioethicists believe that returning this information without explicit involvement of a doctor or genetic counselor is a bad idea. I do not share this view, but I understand it.

• Which brings me to my real take-home message: like most activities in life, personal genomics is not for everyone. I am not terribly interested in bungee jumping or climbing Mount Everest. Were I to let a friend talk me into doing either of those, assuming he wasn’t misleading me, then the onus would be on me for going along with it. But I’ve watched enough YouTube of bungee jumpers and I read Into Thin Air. Ergo, extreme sports and me: not happening. It is incumbent upon potential genome explorers to educate themselves before setting themselves up for possible disappointment and/or anxiety. I worry that critics of consumer genomics start from the assumption that would-be participants don’t know what they’re getting into and ultimately, cannot know, so, you know, let’s call the whole thing off. If you believe, as I do, that genetics education is in a sorry state and must be fixed, then advocating willful ignorance for everyone is both logically and morally indefensible.

• Okay, so who is personal genomics for? People who “opt in” should know upfront that:

• These are still very early days. The information attached to complex conditions controlled by many genes and the environment (like MS) is tentative and subject to radical changes in content and interpretation.
• Some information (e.g., possible high risk of breast cancer) may already be stuff you don’t want to know. “RTFM” as the kids say.
• The ideal personal genomics user is curious about the science, recognizes much of its tentativeness, and is an information seeker such that she won’t shy away from information that may not bode well for her future health.
• Sending a company $400 and spitting in a tube is not the only way to get involved. There are a bevy of research studies that involve personal genomic information. Take some time. See what’s out there, what’s involved, what your rights and responsibilities are, and what your expectations should be. This is your DNA and–unless you’re an identical twin–no one else’s. Do your own due diligence.

• Ask yourself what kind of person you are and why you would want to do this. Despite pressure from friends (and plenty of spam from Facebook), I’ve declined to join Facebook because I don’t like the way FB has handled its privacy policy and the contempt with which it has treated its customers. Plus, I find I already have enough trouble managing my online life; Twitter and email are about all I can handle, not to mention my book pages. Some social interactions make me anxious in a way that my genome does not. Your feelings may be the exact opposite of mine. And that’s perfectly fine: we shouldn’t need each other’s permission.

Personal genomics is, finally, personal. Look in the mirror and decide what’s right for you.

Most of you probably know that Congress, aided and abetted by the Government Accountability Office, came down on consumer genomics companies like a ton of bricks yesterday. I am still sifting through the rubble, but here are some thoughts, questions, Friday morning quarterbacking, and a few provocations:

• Standards. In the course of writing my book, I learned that, for a time, 23andMe, deCODEme and Navigenics (aka The Big Three) were having discussions in an effort to hammer out standards for their services. For whatever reason, those talks fell apart. They are paying the price for that today. The publication of David Ewing Duncan’s genome scan results and the Ng paper, both of which described discrepancies among the companies’ interpretations, were two huge red flags and should have led to 1) a meeting of the minds with a big tent full of experts to try to reach a consensus on interpretation for things like prostate cancer and hypertension; 2) removal of those conditions from their services in favor of, say, clearly useful and less contested things like carrier screening and pharmacogenomics; or 3) a serious downgrade of those results to “provisional” or “preliminary” status. The companies have no one to blame but themselves for these discrepancies, which, in my view, are the most damning aspect of the report and could have been avoided. I would urge them to pursue at least one of the three aforementioned options and get their ducks in a row. Come on y’all: if the oil companies can do it, so can you.
• Ethnicity. Most human genetics research has been done on persons of European ancestry. Genomes of African origin, for example, tend to be more diverse. Thus, with fewer markers typed, consumer genomics test results are apt to be of less value to persons not of European descent. I reckon the companies might have mitigated this criticism by more prominent disclosure of this limitation. Over the longer term, however, the companies are in a position to address this problem–and it is a problem, especially with respect to health disparities–by filling the data gap. We know they can do research, so instead of going to Davos, why not go to Oakland or Tijuana?
• Who are these companies? Dear GAO, if the goal is transparency, I don’t understand why you cannot name names (pdf).  Why be coy? FWIW, here are my guesses: “Company 1″ = 23andMe; “Company 2″ = deCODEme; “Company 3″ = Pathway Genomics; Company 4 = Navigenics. I won’t bore you with my reasoning and I could be wrong, but again, why make us guess? (UPDATE: 23andMe confirms my guesses.)
• Surreptitious testing. Companies: don’t do it. You can sell the spit kit, but no surprise parties.
• The word “diagnostic.” Companies: don’t use it unless you mean it and can prove it. Otherwise you will give Jim Evans a coronary and that would make me sad.
• “You’d be in the high risk of pretty much getting it,” is how a representative responded when our fictitious consumer asked if results indicating she was at above average risk for breast cancer meant she’s definitely getting the disease (see video). If a Navigenics counselor said this when discussing a woman’s risk for non-BRCA1 or BRCA2 breast cancer, then yeah, that’s “horrifying.” But for the sake of completeness, I would like to know the full context of the conversation. What was the quoted risk percentage on the fictitious customer’s profile that would have led her to such a conclusion? I bring this up for two reasons: 1) we’ve seen recently exactly how important such context is with respect to audio-visual evidence, especially when one’s accuser has an axe to grind; and 2) when I had my counseling session with Navigenics, the counselor and I spent much of the time discussing my family history of breast cancer and the possibility I carry a BRCA1 or BRCA2 mutation, despite the fact that Navi doesn’t test for those mutations. But just because the company doesn’t test for it doesn’t mean we can’t talk about it.  These mutations are highly penetrant, i.e., they confer high risk to carriers. Could it be that the counselor was discussing familial breast cancer and not the Navigenics assay? Maybe not, but I would like to hear more of the phone conversation rather than just a brief video with leading statements in the titles.
• Baby meet bathwater. “The test results we received are misleading and of little or no practical use to consumers.” The GAO is entitled to its opinion. Clearly there are issues with these tests. But I found that learning my carrier status for a number of diseases, risk for late-onset Alzheimer’s, and genetic responses to a number of drugs to be useful. I found the ancestry information to be interesting and entertaining. Would I have paid $1000? No. $99? Yes. More than that? It depends on what I’d be getting. That’s why the GAO’s failure to distinguish among the companies, their pricing structures and their offerings is so unfortunate. The net result does not allow for any nuance whatsoever. Rather, if one accepts the GAO’s conclusions, there seems to be room for only one response: “Off with their heads.” If that’s the case, then consumer genomics will be driven underground and/or overseas. In my view, that would be a shame.

We saw a very similar hearing — with the same undercover M.O. and public perp walk — in 2006. Were/are some companies behaving badly? Absolutely. And they should be called on the carpet. But let’s not kid ourselves: this dog and pony show was as much an opportunity for chest beating and righteous indignation as it was for “protecting the public health.” It was an opportunity for politicians to scream “snake oil” and fret about the possibility of their constituents “jumping off a building” as much as it was an attempt to bring about constructive change to a nascent industry’s practices. Why is that? Why is the demand for data never made of the alarmists? Why do the companies not get to see the GAO’s report ahead of time? Why did the FDA exhibit a genuine spirit of open inquiry and invite a broad range of stakeholders to its meeting while Congress could not be bothered to do the same? Why does the outcome of such hearings always seem to be a foregone conclusion?

Representative Stupak says he and his colleagues are on top of their game.

What game is that exactly?

One woman panicked when the genetic test she had ordered over the Internet concluded that her son was carrying a life-threatening disorder and, even more disturbing, that he was not — genetically — her son. Another, who always thought she was white, was flabbergasted to find her genes were mostly of African origin. A third woman’s result was still more stunning: She was a man, it said.

“I thought, ‘Oh my God. Am I really a man?’ ” said Denise Weinrich, 48, of St. Peters, Mo. “I thought, ‘What’s the matter with me? I’m not who I thought I was. How am I going to tell my children?’ DNA doesn’t lie.”

Ahem.

Daniel MacArthur from Genetic Future, Dan Vorhaus from Genomics Law Report, Luke Jostins from Genetic Inference, and Caroline Wright from the PHG Foundation have launched Genomes Unzipped:

Our goal is to provide you with independent analysis of advances in the field of genetics, with a particular focus on implications for the budding industry of personal genomics. We’ll also be discussing ways in which you can make the most of your own genetic data using online resources and techniques developed by researchers.

My data are in the mail, y’all.

In an effort to stop DNA testing on state campuses, a bill was introduced in the California legislature last week that would prevent the University of California system from:

“making an unsolicited request to an enrolled or prospective student of that segment for a DNA sample for the purpose of genetic testing.” It would also require that the universities report how much they are spending on such tests; the schools would then have their funding reduced by that amount.

Keep moving those goalposts, boys and girls. We may not be able to legislate morality, but ignorance is clearly within our grasp.

(hat tip: Dan Vorhaus)

This month, PLoS Genetics is publishing an article from the company 23andMe reporting the first genome-wide association studies (GWAS) on multiple traits ascertained by self-reported information provided through the Internet from over 10,000 participants who pay the company for providing whole genome genotypes [1]. The paper passed through scientific review by a panel of three experts relatively quickly and is sure to attract the attention of anyone with freckles, curly hair, or an aversion to asparagus. Novel associations are described for four intrinsically interesting traits (out of 22 considered), while known associations with hair and eye color are replicated in a dynamic data-gathering context. Additionally, intriguing observations on the interaction between genetic self-knowledge and self-report of phenotypes are described. The implications of the successful application of this Internet-enabled approach to GWAS research were considered to be more than sufficient to warrant publication in the journal.

 

***

The editors of PLoS Genetics recognize that the decision to publish this study, without IRB review as human subjects research and with some concerns over the consent document, and the fact that there is limited access to the raw data, will not sit well with some, perhaps many, readers. As outlined above, though, a prima facie valid IRB exemption was obtained, and, while there are ambiguities in the consent form, there was no evidence that these amount to an inadequate document. After considering all of the evidence, we decided that publication, accompanied by an editorial providing transparent documentation of the process of consideration, was the most appropriate course.

Seems like this Interwebs thing might be starting to catch on.

(Hat tip: the indefatigable Bora Z) 

I have a piece in the June 23, 2010 edition of Slate (is it fair to call it an edition?) on the recent kerfuffle surrounding personal genomic testing on Bay Area college campuses:

A new kind of testing is coming to university campuses across the country. In the last month or so, Berkeley and Stanford each announced that their students would be offered the opportunity to have their DNA analyzed from saliva samples. The range of reactions to this news—from horror at the thought of young people misinterpreting their results to unbridled enthusiasm at the chance for large-scale genetics education—confirms that the field of personal genomics continues to mystify. Genetic testing for disease has been around for 50 years now, yet we have no consensus on how and even whether it should be used in healthy people.