So here we are again.
Last year, a Congressional panel met to discuss consumer genetic testing. Some of the speakers issued statements that were so factually inaccurate as to be offensive; then, they largely concluded that people should be protected from their own curiosity. This week, an FDA panel met to discuss consumer genetic testing. Some of the speakers issued statements that were factually inaccurate and in one case plain offensive; then, they largely concluded that people should be protected from their own curiosity.
When I wrote a six-part series about consumer genetic testing last year for Newsweek, I was not an advocate on either side. (At least I didn’t start out that way.) Now I find myself squarely in one camp and sorely disappointed in the way the dialogue on the issue has played out over the last few months. What motivated me to write about the issue in the first place was a feeling that consumers were being left out of a debate that was supposed to be about them. Critics in Congress and the GAO were concerned that people would mistakenly treat their necessarily incomplete genetic data sets as crystal balls, but how many customers were actually invited to the Congressional hearing to share how they might think about their results? None. This time around, at the public meeting, how many consumers were expressly invited to speak? None.
It’s hard to see how people in power could ever look fairly at consumer genetic testing, given that every time they tackle the issue they end up having the same conversation without bothering to get direct input from the people whose welfare they’re purporting to protect. UPDATE: After the meeting, the FDA held an official public comment period. This was a great way to to get consumer input, and the agency should be commended for it.
So here I am again, waving my hands and jumping up and down, trying to provide some of that input.
Last year, I was a potential consumer of genetic scans. Now I’m a current consumer. I bought a kit from 23andMe in August; when I got my SNP list back, I sent it off to deCODEme for a generously donated second opinion. And — attention, FDA! over here! — I found out something earth-shaking.
It wasn’t that I’m at dramatic risk of Alzheimer’s, which was the only bad news I thought I might get. Turns out that 23andMe doesn’t test for variants, or SNPs, in APOE4, which is the gene that could have told me such a thing. Scientists can often figure out what a genetic variant is without directly testing it, by examining other nearby variants — this is called imputation — but for technical reasons, APOE4 data apparently can’t be imputed from the information I have. The upshot is that if I’m carrying a variant that might dramatically increase my risk of Alzheimer’s Disease — an illness to which several of my relatives have succumbed — I have no way of knowing, at least not from these results.
It wasn’t that I have a very high or low genetic risk for any other common disease. Best I can tell from these results, I don’t. I suspected that already based on family history. My SNPs are just as “delightfully uninteresting” from a medical perspective as I expected them to be.
I did have two health risks that appeared in bold red text at the top of my 23andMe results list. I suppose that bold red text might have scared me if I hadn’t read the thin black text that was intended to explain it. But I had paid $500 for the thin black text. So I read it, and I reread it, and I cross-referenced it with my deCODEme report, and I clicked on the links to studies, and I did some literature searches of my own. Here is what I concluded:
(1) For the first bold red risk — atrial fibrillation — the results were based on two SNPs. Neither by itself correlates with a dramatic rise in susceptibility. (Also, I have only one of the high-risk SNPs; my other associated SNP is actually linked to a slightly lower risk than average.) Both SNPs are in the same gene-poor region of chromosome four. Nobody knows what they do, biologically. One thing scientists do know is that other genes have been and will be linked to atrial fibrillation. None of these other genes were tested in my results. If they had been, who knows what would have shown up? It’s possible I’m not at high risk for this disease at all, especially since I have none of the known environmental risk factors. In short: I’m not terribly worried.
(2) The second bold red risk was for Restless Legs Syndrome. According to 23andMe, 4.2 out of every 100 European women develop RLS; among those with a particular variant on chromosome six, which I have, 5.2 out of 100 will. (This in and of itself made me skeptical. More than four percent of European women have this disease? I’m willing to believe the illness is real, but might it be overdiagnosed? Why, yes, it might be.) Unlike my atrial fibrilliation SNPs, this variant is in a gene, albeit in a non-coding portion. 23andMe does an excellent job of explaining what that may mean: “The function of this gene is not yet known. Scientists do know that it belongs to a large family of genes that encode proteins that can influence activity levels of other genes. The SNP in this gene doesn’t actually cause a change in the protein sequence of BTBD9. Instead, it lies in a non-coding part of the gene where it may affect how BTBD9 is turned on or off.” In other words, I have one variant that may help regulate one gene that regulates other genes; this variant is linked to a polygenic disease for which the background rate is in dispute. Gosh! Forget Restless Legs Syndrome; isn’t this knowledge going to keep me up at night? No. No, it’s not.
These were the only two higher-than-average risks in my report. My analysis of the rest of my risks proceeded along the same lines — which were lines that anyone with an ounce of sense would follow. To wit:
Hey, I have a lower-than-average genetic risk of melanoma! Except I might not, seeing as how my 23andMe results are based on just two SNPs (does anyone think cancer is caused by two genes alone?). Given that there is melanoma in my family, and given that I’m a white girl who regrettably got burned to a crisp on a regular basis in childhood, I’ll keep wearing sunscreen.
This just in: my AA status at the SNP rs7590720 suggests I’m at “typical odds of alcohol dependence in men!” Except I might not be in real life, because alcohol dependence is not a Mendelian trait and I’m not a man.
Cool, I’m AA at rs363050! According to 23andMe, a 2006 candidate-gene study of 667 Dutch subjects showed that for each A at this site, non-verbal IQ goes up three points. But candidate-gene studies often fail to replicate. (Small quibble with 23andMe: Why cite candidate-gene studies on such a complex trait, especially in the era of GWAS? Maybe an appropriate role for regulation would be quality control, assuring that the literature cited is sound.) Besides, IQ is presumably affected by all kinds of things. This result is useful only for making jokes about one’s prowess at Sudoku, and one doesn’t need to be AA at rs363050 to see why.
What else? I’m slightly more sensitive to Coumadin than average and slightly less sensitive to Plavix. I’ll bear that in mind if I ever need an anti-coagulant — but for now, I’ll content myself with learning more about how anti-coagulants work. There’s some fascinating stuff there, starting with the fact that Coumadin was invented as a rat poison.
I carry none of the rare recessive variants that I got tested for, except for one linked to hemochromatosis. Were this a disease-causing variant, I might ask my husband to get tested, too, but my variant is mild. The only consequence of my carrying it, and finding out that I do, is that now I’m interested in hemochromatosis.
I had a blast figuring all this out. And I didn’t need a doctor to help me do it. Actually, sitting in a doctor’s office trying to sift through the data would have been less fun and almost certainly impossible, as I’m sure my primary care provider has much better things to do than babysit me while I look at PubMed.
Now that I have my results, I can’t help but laugh at myself a little, or at least at the version of myself I used to be. Listen to me back in August, agonizing over whether I really wanted to check out my own DNA: “When I finally pulled my testing kit out of its box, [I] found to my complete surprise that I was — there’s just no other word — scared of it. … Whether or not the test qualifies as a medical device under FDA rules, it looks like one. It has a biohazard sticker on it!”
Why on earth was I so scared? What was it that caused me to see my own spit as a biohazard? If you had asked me back then, I might have said something like “emotion” or “irrationality.” Now, I know that this particular type of irrationality has a name: “genetic exceptionalism.” It’s defined as “the idea that genetics must be treated as special under the law,” and it’s largely based on the unfounded belief that DNA is more powerful than it actually is, to the point of being magical.
I don’t know why genetic exceptionalism is so common and reflexive. I don’t know why so many people seem to naturally approach DNA with fear and awe. Maybe it’s because almost everyone is introduced to genetics through Mendel’s pea genes, which really were individually powerful in their effects on phenotype; maybe the lesson people subconsciously absorb is “all genes work like that.” Or maybe it’s just because of GATTACA.
Maybe it’s because DNA really is special in a number of ways. It is a biochemical form of literature, a Great Natural Novel that tells a story of life. (Not the whole story, by any means. But a good story.) It is elegant and beautiful.
But it is not magical or all-powerful, and if we are ever going to make responsible policies around it, we have to stop thinking of it as such.
Here is the earth-shaking thing I found out: It is possible to stop thinking that way. Reason can overcome emotion. People can approach information about DNA, even their own DNA — information that they might naturally find scary — and come to realize that there is nothing to fear because they see it in cold clear light for nothing more than what it is. I know people can do this, because it’s what I did.
Genetic exceptionalism is the default, the state of ignorance. How do you combat ignorance? By enabling knowledge. By giving people information and showing them what it may mean for their lives. By, say, appealing to people’s inherent curiosity about themselves, and then handing them a document that shows — that cannot help but show — that genes aren’t destiny.
This is precisely the information that bad regulation would keep out of the hands of consumers.
What will happen if we, as a society, confirm people’s misconception that genes are all-powerful, and if we tell them they are too stupid to understand their own data? They will come to believe that genes are all-powerful and that they are stupid. It’s a self-fulfilling prophecy. Whatever you think of consumer genetic testing, there is nothing that can turn up in a scan of a few hundred thousand SNPs that is anywhere near so frightening.