Genomes, cool conferences, and what the hell to tell people about behavioral genes

I had the pleasure of attending the Genomes, Environment, and Traits conference on Tuesday. Was wonderful and strange, with many inspiring, exciting, and/or entertaining moments -- and a few things a bit worrisome.   

The twitter feed from the event tracks the talks and agenda pretty thoroughly; it's far better than my own notes. I especially enjoyed the morning's main event, in which a tag team of Robert Krulwich and Carl Zimmer called to stage for interviews different combinations of 13 the 10 "pioneers" who had been among the first to have their entire genomes run. As a journalist, I had wondered how Zimmer and Krulwich would handle this. The program said they would interview all 13 ten. When I pictured all of these people lined up on stage while the two journalists tried to interview them, I pictured chaos and trouble. These pioneers featured some larger-than-life figures and a couple cowboys, including James Watson and Skip Gates: a diversity and range of energy, personality, and ideas that might prove hard to handle.

But Krulwich and Zimmer had stayed up late forging a plan that proved up to the task: divide and conquer. They brought up Watson first, and alone, and extracted from that dangerously unpredictable and sometimes caustic presence 15 minutes of humorous and insightful history and charm. Then they brought up the other genome'd pioneers -- Jay Flatley, Esther Dyson, Stephen Quake, Misha Angrist,  James Lupski, Seong-Jin Kim, Greg Lucier , Rosalynn Gill, and the West Family--  in various combinations to address various business, personal, historical or ethical angles in friendly but often pointed conversations.

It worked beautifully. A highlight was Skip Gates. Gates talked about how getting his own genome done seemed to bring his mother right into the room. Then, talking about his TV and DVD series on genomic geneology, Faces of America, he first slayed with an hilarious account of how he used his friendship with Quincy Jones ("Quince, what's up?") to draw a coveted phone call from Oprah Winfrey, who convinced the Coca-Cola company "to split open the ceiling and lower a giant ATM machine that just started spitting out money"; and then moved us by describing how the overlapping genetic heritages he looked at in the series -- personal geneologies of whites, blacks, Asians, Native North Americans (the incomparable Louise Erdrich), and others -- crossed and joined one another in ways so rich and unexpected that they "completely destroyed any notion of racial purity" and "showed we're all a great big boulliabaise." Beautiful.

And splendid programming and interviewing from Krulwich and Zimmer, especially with Watson and Gates. Those two you don't know what you'll get -- might be genius, could be trouble. But from each they got 20 minutes of their most inspiring and engaging stuff.

The other interview guests and combinations held their own too, with many good tales, surprises, and laughs, and an almost frighteningly intelligent and self-possessed 17-year-old.

But it wasn't all charm and entertainment. We talked a lot all day about balancing the promise of genomics -- reliable insights that would help people understand themselves and manage their lives and medical care -- with the many problems in putting it to use, such as inevitable gaps in knowledge or execution. A major concern was how to analyze the information and help people understand what it means. As many there noted, we can now produce a person's raw genomic data pretty quickly -- but we aren't nearly as fast or clear in discerning its real significance. Some genetic analysis companies are working hard to do this. Others are focused mainly on pumping out the raw data. And whatever the intentions and plans of the most responsible players, the overall move here, as in so many fields, is toward producing and selling data at cheaper and cheaper rates to a broader array of people. Some of these companies may do a bangup job of explaining the results. Others, maybe not.

And so emerged through the day (to me, anyway) a sharp and deep tension between the genuine potential to use genomic information to improve medical and other personal decisions on one hand, and on the other, the extreme difficulty of doing so when even the medical and genetic worlds don't know or agree on what much of this information means â and the industry is racing to put this information directly into the hands of people who have little idea what to make of it.

Dan MacArthur had a good post today on a new paper that shows the technical difficulties simply extracting the genomic information. Interpreting it may prove even trickier, particularly with genes related to behavior or mood. As I'm quite interested in those, I couldn't help but notice that they didn't come up a lot in the formal discussions. But when I talked to people on the side, including some of those who had their genomes run, they usually confirmed my impression that people take a particularly keen intereste in genes related to things like mental health or behavior -- depression, bipolar, hyperactivty, aggression. "Oh God yes," one person told me. "Unless you're really worried about cancer or something, that's the first thing people look at. 'Do I have the crazy gene?'" Yet by my read, neither the industry nor the research community quite knows what to tell people to do with that information -- even as we move closer to making it cheaply available.

We discussed this in the breakout session I led, on "Predicting Temperament." Predictably enough, we didn't come up with all the answers. I'm not sure what they are. Cheap genomes are good. Information wants to be free. You can't stop these things. But it seems to me that people in both genomics and behavioral science need to talk a lot more about how to help people make sense of not just the classic medical implications about things like cancer, aging, or diet, which are tricky enough, but about behavior. Many people likely to be tested take a keen interest in these genes. And well they should, since temperament, mood, and behavior drive so much of our fates. Yet the testing, genomics, and behavior communities don't seem to know what to tell them. We talked about that some at GET; kudos to George Church, Jason Bobe, Tom Goetz, and the other organizers for making that possible. We need to talk about it some more.

PS April 30: Zimmer has put up a nice post on the conference, as has Tech Review's Emily Singer. And Zimmer called out a piece I'd forgotten about, his own, about the slippery business of typing genes for intelligence. Also notable (via Zimmer again) are Steven Pinker's long 2009 article in the Times Magazine and George Church's "Genomes for All" manifesto [pdf] from from Scientific American. And Wired editor Tom Goetz has a 2007 article on his own and others' genotype data.

*Apologies on this front to readers. I've been traveling, conferencing, and researching at a furious pace lately and so have been especially short on blogging time. This pace should soon slow, however, leaving me more time to post on genes, behavior, and much else.

More like this

The Genomes Environments Traits conference in Boston is without a doubt the place to be on April 27th for anyone interested in personal genomics: the conference has managed to attract nearly every human being in the world who has had their complete genome sequenced (excluding, of course, anonymous…
Update: In the comments below, SNPedia co-founder Michael Cariaso notes that Duncan has already lost his crown to the anonymous European NA07022, recently sequenced by Complete Genomics, who weighs in with 5891 associations to Duncan's 5321. Records don't last long in the age of high-throughput…
A few calendar notes: I've got a three-day run starting next Sunday in which I'll be talking to authors and journalists about book proposals; NY science writers about the future of social media; and to genomic geeks about genes and temperament. If you've questions you'd like raised at any of these…
Lupski, J.R., et al. (2010). Whole-genome sequencing in a patient with Charcot-Marie-Tooth neuropathy. New England Journal of Medicine advance online 10.1056/nejmoa0908094 Roach, J.C., & et al. (2010). Analysis of genetic inheritance in a family quartet by whole-genome sequencing. Science : 10…

Sounds like a fantastic conference. We are still very very far from being able to predict even major disease risk accurately from a single genome, however - mainly because we have not yet found most of the genes that are involved (say in psychiatric illness, for example). And predicting behavioural traits or temperament will also be very difficult on an individual basis - trends across populations due to particular genetic variants may emerge but applying these in the context of a unique genome will be hugely problematic. More fundamentally, though, is the fact that much of the variance (usually at least half) in these traits is not genetic. So there are major limits to how much genetic information can tell us - one of the strongest factors limiting genetic determinism is simply intrinsic developmental variation - even if you start with the same genotype, you never end up with exactly the same phenotype.

See this post on Wiring the Brain: "Nature, nurture and noise" for more on this topic: http://wiringthebrain.blogspot.com/2009/06/nature-nurture-and-noise.html