A couple of science-writing colleagues objected to my recent post on the "warrior gene." Both accuse me of throwing the baby of modern behavioral genetics—which seeks to link complex behavioral traits to specific genes—out with the bathwater of media hype. Naturally, my innate bellicosity compels me to respond.
My warrior-gene post examined the reported association between aggression and MAOA-L, a variant of a gene that helps produce the enzyme monoamine oxidase-A. The journal Science dubbed MAOA-L the "warrior gene" back in 2004 and the name stuck. Ed Yong, the British blogger, says in a comment to my column that just because some folks sensationalize the MAOA-aggression link doesn't mean it's bogus.
Ed explains that "children who carry MAOA-L AND come from abusive homes have a higher risk of aggressive behavior; that's not true if they come from more stable backgrounds. Likewise, the hot sauce experiment found that people with MAOA-L are more likely to mete out punishment when they are provoked—another case of nature via nurture. It's sad that the fascinating area of gene-environment interactions isn't discussed at all" in my post.
Ed focused on these "gene-environment interactions" in his 2010 article, "Dangerous DNA: The Truth about the 'Warrior Gene.'" Ed cautioned that MAOA-L "is not a gene 'for' aggression," but he accepted the notion that MAOA-L plus certain environmental stimuli raises the risk of aggression. First of all, the hot-sauce study found a minute difference, at best, between the MAOA-L subjects and those carrying the more common MAOA gene. That's why I didn't find it credible.
Ed also touted a 2002 report in Science by Avshalom Caspi of King's College, London, and seven colleagues that MAOA-L carriers were more likely than non-carriers to display antisocial behavior, but only if they were "maltreated" as children. What Ed doesn't say is that two studies by geneticists at the University of Colorado have failed to confirm the Caspi claim; you can read these reports here and here. Did they get any attention? Of course not. This has been the pattern of media coverage of behavioral genetics since I started following it in the late 1980s: The spectacular claim makes headlines, and the counter-evidence doesn't. The public and even many journalists and scientists are left with an erroneous impression that behavioral genetics is explaining more and more of what we do.
I'll call the other journalist who knocked me, who prefers to remain nameless, Ed2. He accused me of pulling a "bait and switch," in which I conflated "pop misuses of the 'warrior gene'" with the overall performance of behavioral genetics, which deserves more respect. When I asked him for examples of robust linkages of specific genes to complex behavioral traits, Ed2 gave me two.
One is research linking Lesch-Nyhan disease, a rare disorder affecting one out of every 380,000 people, to a gene called HPRT. Lesch-Nyhan causes over-production of uric acid, which results in painful buildup of uric acid crystals in joints, the urinary system and subcutaneous tissue. Sufferers sometimes strike and bite themselves, even gnawing off their own fingers. Lesch-Nyhan is clearly not a behavioral disorder but a physiological disease with behavioral side effects. Ed2 was pulling a bait and switch on me!
Ed2's other example is a 2010 paper linking the gene CMYA5 to schizophrenia, which unlike Lesch-Nyhan has no definitive physiological symptoms and so counts as a complex behavioral disorder. The CMYA5 study looks impressive; it involved a huge team of researchers from dozens of institutions around the world examining more than 33,000 subjects. Reading the data-dense paper, I couldn't figure out how much CMYA5 supposedly increases the risk of schizophrenia. I emailed the lead author, Sam Chen of Virginia Commonwealth University, and he told me that if you have the CMYA5 gene, your risk of schizophrenia increases by seven percent.
Let's put this statistic in context. About one in 100 adults around the world are schizophrenic, which means that the risk of schizophrenia for the general population is 1 percent. If you carry the CMYA5 gene, your risk rises by an extra 0.07 percent to 1.07 percent, according to Chen. If you have a schizophrenic first-degree relative, such as a sibling, your probability of becoming schizophrenic is about 10 percent, which is more than 100 times the added risk of having the CMYA5 gene. My guess is that the finding of Chen et al. will eventually be discarded as just another false positive, like all the other alleged "schizophrenia genes" dating back more than two decades. But if even if the correlation holds up, what good is it? Is research like this really worth the effort?
When it comes to behavioral genetics, so far there is no baby; there's only bathwater.
John Horgan, Scientific American