Your genetic info -- not free, easy, or clear

After I wrote in my Atlantic article about getting my serotonin transporter gene assayed (which revealed that I carry that gene's apparently more plastic short-short form), I started getting a lot of email â several a week â from readers asking how to have their SERT gene tested. This led to an interesting hunt.

It was a hard question to answer. I couldn't just tell people to do what I did, for a psychiatric researcher/MD I'd known for years, who specializes in depression and serotonin, had done mine as a sort of favor to science and journalism. That researcher also stood by, had I needed it, to offer counseling and more information about the result's implications â an important point.

Obviously I couldn't pass that researcher's name out to several dozen strangers. Yet the readers who wrote wanted the information for the same reason I did: They wanted to know whether they had a genetic variant that by the conventional reading simply increased your risk of depression, but by another reading â the hypothesis explored in my article â conferred a broader sensitivity to experience, which can be a good thing.

As I well knew, this is powerful information. It opens a big box with lots of compartments. Those compartments hold things like the recently raised (but far from settled) question of how solid a connection the S/S variant holds to depression. They hold the question of how a person might interpret the results, and whether they have the emotional, intellectual, and social assets to make the most of the information; as well as the whole pile of broader issues raised by the growing availability of spitomics.

Challenging territory. Yet in this case the biggest challenge my serotonin-curious readers encountered was just getting the test in the first place. For try as we might, we could not find a place to run a SERT assay. People would try various companies, such as 23andme, and report back that they'd struck out. I suspect this is because the lawyers at such companies likely advise against giving test results regarding mental illness to people whose very interest in the test suggested they might be prone to depression.

So I started suggesting the long route to my solution: Try to find a psychiatrist at a research university interested in serotonin and convince that person, with your doctor's help if need be, to run the assay. That too proved elusive.

But one particular reader (I'll call her Natalie, which is not her real name) refused to give up. She wrote me back in early February saying her therapist had given her the Atlantic article, and she wanted to know her SERT type. She suspected she was indeed S/S. So off she went a-hunting. She struck out with the major gene-testing services, but after several back-and-forths and dead ends, seemed very close to getting an assay done at a nearby research university. The psychiatrist there, having read the article, seemed game, and set up an appointment. When she visited him, medical history in hand, the psychiatrist "was conscientious, thorough, and very interested, but understandably cautious." She liked him. But he ultimately declined to prescribe the test. He told her, she wrote me, that

the general body of research was inconclusive and contradictory, [and] professional organizations (such as AMA for geneticists) advised against it. Presently doing the test "would do more harm than good" and I should check back every few years when perhaps the test would provide concrete information. I think he was highly ethical and also covering his ass.

I think she's right. It's easy to understand this doctor's response. Some have lately questioned whether the S/S confers risk. Even if you believe the S/S does confer risk (and the field is very much split on this), do you want to deliver that news to someone whose depression might lead them to take the news badly? The view of S/S as a 'risk' gene rather than a plasticity gene remains the prevailing paradigm. And it seemed to drive the no-test decision Natalie and other readers encountered âeven though, in Natalie's case, she had shown the researcher the article and told him she was viewing the potential results as a sign of sensitivity rather than just risk.

But just at this point, where the road seemed to peter out completely, Karen found, hidden in the weeds, a lab that would run the SERT assay. Why would they do it?. Because, they said, "Researchers have found that mutations in some genes specifically change how effectively SSRIs may act," and that in the SERT gene, "[p]patients who have the short allele are less likely to respond to SSRIs or may take longer to respond."

This is true â but only in some trials, but not in others. The data on the SERT gene's effect on SSRI response is mixed. Yet it's interesting that the company lean on this to offer the test.

No matter to Natalie. She's primed to see what she's got. "I can tell you right now I am hoping for the short-short," she says. "The process is in the works for me, so now we shall see. This is so much fun."

I obviously think people should be able to get this information â though they should get it with some informed counseling. But this entire hunt, and the various reasons to give or not give people this information about themselves, raises a mess of intriguing and often slippery issues. The genetic testing industry, and the understandable excitement about the potential and power of genomic information, generally assumes that we either don't know a gene's meaning or that we do. But the case of the SERT gene shows that a gene's meaning can be far from certain -- and that it can change substantially as researchers do more work on it and view the existing data from different angles. Here we have a gene variant â possibly the most prominent in behavioral science, certainly in psychiatry â that, depending on whom you ask, is bad news, no news, or news whose meaning ... depends.

As the genome's parts come into view, their meaning sometimes shift. How will the fast-spreading, fast-growing gene-information industry handle this movement? How do you give people advice about a gene whose meaning seems to be changing? How, say, would a pre-conception gene counseling service, something along the lines of Counsyl, handle this dilemma?

This isn't something we'll figure out in a few blog posts; it's something the industry and the broader genomics community will need to consider carefully over the next few years, even as it rapidly grows. I'll be talking about and leading discussions on these questions at the upcoming GET conference in Cambridge and then at a closed workshop at the Institute for the Future â a start. Your own ideas â as well as pointers to other explorations of these questions â are most welcome in the comments or via the Twittersphere.