The pith: there are differences between populations on genes which result in “novelty seeking.” These differences can be traced to migration out of Africa, and can’t be explained as an artifact of random genetic drift.
I’m not going to lie, when I first saw the headline “Out of Africa migration selected novelty-seeking genes”, I was a little worried. My immediate assumption was that a new paper on correlations between dopamine receptor genes, behavior genetics, and geographical variation had some out. I was right! But my worry was motivated by the fact that this would just be another in a long line of research which pushed the same result without adding anything new to the body of evidence. Let me be clear: there are decades of very robust evidence that much of the variation in human behavior we see around us is heritable. That the variation in our psychological dispositions, from intelligence to schizophrenia, is substantially explained by who our biological parents are. This is clear when you look at adoption studies which show a strong concordance between biological parents and biological children on many metrics as adults, as opposed to the parents who raised the children. This doesn’t mean that environment doesn’t matter, but I believe we tend to underweight genetics in individual outcomes in our contemporary Zeitgeist, just as we may have overweighted it in the past.
At this point some of you may be wondering, “what, I hear about genes for [fill in the blank] constantly!” So why am I saying we underweight genetics? I think there’s a disjunction between the fixation that the public (and therefore the popular press) has on a specific biophysical candidate gene which is given almost magical powers of causal necessity and the more abstract and diffuse statistical genetic reality of correlations between parents and offspring whose effects seem to be distributed diffusely across the genome. The latter is a robust and ubiquitous phenomenon, but because it is not possible to frame the narrative as a “gene for X” it lacks power. In contrast, when you have a powerful gene of large effect whose variation in state has a concrete and comprehensible outcome the narrative is clear, precise, and distinct. There’s an unfortunate problem with this though: quite often the narrative is wrong because it is not robust. It won’t be replicated and stand the test of time.
The example of the “language gene,” FOXP2, is illustrative of the issues I’m pointing to in the most broad of terms. As a matter of fact FOXP2 is a much better candidate for being the “language gene” than is usually the case for the gene for X, but ultimately it is probably not alway useful to term FOXP2 the language gene when the faculty for speech is such a complex trait subject to many biological pathways. The putative “God gene” was a much worse case of a gene for X, and probably a good example of the problem I’m talking about. There’s a pretty robust body of evidence that religiosity has a heritable component, but there isn’t much evidence for a gene for belief in God.
What does all this have to do with dopamine receptors and novelty? The DRD4 locus has been implicated in a lot of behavior genetic variation, and dopamine receptor genes are often pointed to as “master controllers” of a sort for various aspects of personality and life outcomes. Dopamine as a neurochemical has myriad functions, so variation in its production controlled by genetics is a natural candidate of interest for researchers. The problem is that the nature of this sort of statistical and sexy science is that there’s going to be a natural gravitation toward significant results which later turn out to be false positives. Before moving on, I do want to reiterate that as a gene for X the dopamine receptor loci are much better set of candidates than the “God gene,” but here the devil is in the details.
Let’s see what the argument in the paper which triggered The New Scientist piece is. Novelty-seeking DRD4 polymorphisms are associated with human migration distance out-of-Africa after controlling for neutral population gene structure:
Dr. Daniel MacArthur points to a long article by Edmund Yong, Dangerous DNA: The truth about the ‘warrior gene’. Dr. MacArthur notes on his twitter account: “Nice piece on behavioural genetics…but should emphasize MOST behav. gene assocs are actually false.” I think he’s pointing to the winner’s curse; there are lots of people studying various topics, but only a subset of studies pass which yield appropriate effect sizes and p-values actually get published. A sequence of such may give a false sense of certitude as to the strength of the association between a locus and a trait, as negative results are not usually published. I hope David Dobbs keeps this in mind in relation to his new book on the ‘orchid hypothesis’. We have decades of research which suggest that a lot of human behavior is due to variation in genes within the population. In other words, many psychological traits and predispositions are heritable. But both the earlier linkage studies and now the associations which try and establish a particular gene as the primary causal factor are much more provisional, and like much of science wrong or ultimately of marginal long term value.
The incredible amount of press which genetics and genomics research with behavioral implications receive in the press is more about our psychology than the state of science as it is now. Similarly, consider the enormous swell of neuroimaging research within the past decade. Both genetics and neuroscience offer up the possibility of establishing a sturdier biophysical grounding for the human sciences, but we shouldn’t get ahead of ourselves. Finally, the fact that we know that psychological traits are heritable is useful in and of itself, whether we know the underlying genetic architecture of the trait or the neurobiology mediating between the genetic and behavioral level. Look to the parents, and you shall know a great deal.