Your impatience is in your genes! (well, some of it)


Nature Neuroscience has a short communication which is very intriguging, Genome-wide association study of delay discounting in 23,217 adult research participants of European ancestry. How’d they get such a large sample size? Collaborating with our friends at 23andMe.

That being said, the abstract leaves a little to be desired:

Delay discounting (DD), the tendency to discount the value of delayed versus current rewards, is elevated in a constellation of diseases and behavioral conditions. We performed a genome-wide association study of DD using 23,127 research participants of European ancestry. The most significantly associated single-nucleotide polymorphism was rs6528024 (P = 2.40 × 10−8), which is located in an intron of the gene GPM6B. We also showed that 12% of the variance in DD was accounted for by genotype and that the genetic signature of DD overlapped with attention-deficit/hyperactivity disorder, schizophrenia, major depression, smoking, personality, cognition and body weight.

First, “Delay discounting (DD)”, is another way to say that you have high time preference. That is, you won’t forgo some gains in the short term for greater gains in the long term. You would really “fail” the marshmallow test.

Though there have been legitimate criticisms of the replicability of the effect size of the marshmallow test, there almost certainly is something to time preference and delayed gratification, and its relationship to the ability of young children to master the marshmallow test. In a macroeconomic sense societies characterized by low time preference can sustain lower interest rates, and lower interest rates have all sorts of stimulative properties on long-term economic growth.

But to be clear, the paper above does not detect a variant SNP, rs6528024, which explains 12% of the variance in DD. Rather, 12% of the variance could be accounted for by SNP-chip variance. That is, one could explain the “missing heritability” using the markers they had. The total heritablity of the trait is quite higher, 46% to 62% proportions are citied in the paper (narrow-sense). This means that of the total variance of the trait about half could be explained by additive genetic variance. Obviously the SNP-chip only captured a small minority of that additive genetic variance.

DD is correlated with a lot of things. There is a positive phenotypic correlation with:

  • Smoking
  • Substance abuse
  • Obesity
  • ADHD

They observed a positive genetic correlation between the variants associated with DD and:

  • Smoking
  • Neuroticism
  • Depression

And a negative genetic correlation with:

  • College completion
  • Years of education
  • Childhood IQ
  • Schizophrenia

In relation to the last, schizophrenia and DD are positively correlated phenotypically. That probably means that the underlying genetic causes of schizophrenia and DD are very different.

The patterns of correlations offer up a lot of avenues to speculate. They do a little of it in the paper, but are appropriately cautious. It seems entirely likely that in the near future we’ll be able to characterize a lot of the heriability genomically. When we figure out time preference and intelligence we’ll have come close to answering many of the questions that explain why different people have different life outcomes.

Note: It is no surprise that there is a negative correlation between DD (high time preference) and conscientiousness. Also, the association they found, GPM6B, has pretty clear biological relevance. It’s almost certainly real.

Liberals will never disappear (neither will atheists)

In Quillette Hrishikesh Joshi and Jonny Anomaly* ask Are Liberals Dying Out? Since the piece has been shared a fair amount (judging by my Twitter timeline), I thought I should respond to why I don’t think that is a major concern. Let me jump to their last paragraph:

Nevertheless, despite cultural trends, the best available evidence suggests that political ideology is heritable, and that people with liberal personality traits currently have far fewer children than conservatives. If this trend continues, it is possible that the reproductive choices people are making today will influence the political climate of future generations. Over the long run, conservatives could end up winning the ideological contest with fertility rather than arguments.

First, I don’t think the title reflects the modest contentions of the piece. I beseech the editors of Quillette to not engage in the titular hyperbole so common in the mainstream media!

I agree that political orientation seems heritable. That is clear in books like Born That Way. But heritability expresses itself in an environmental context. If you had a totalitarian government most of the phenotypic variation would disappear. Yes, there would be dissidents, but they’d be freaks. Most humans would conform (no, I don’t think the citizens of Soviet Russia were genetic freaks unable to grasp freedom like Howard Roark). The correlation between religiosity and fertility varies by society as well. The more secular the society, the bigger the gap (though last I checked this was not true in China). In a totally conservative future heritable variation for liberalism could just reemerge.

Second, political orientation exists on a relative plane. If one imagines it as some specific thing, or disposition, one can imagine that in the future the liberal-conservative spectrum would exist, but just be shifted. Quantitative genetics has shown that selection can move the mean many standard deviations. I don’t think this is a strong objection to their overall point, but it gets at the fact that we view liberal-conservative tendencies along a distribution (1980s liberal commentator Jeff Greenfield was widely known for making disparaging comments about gays i the prime of his career; that did not destroy his career as a liberal pundit at that time). Perhaps liberal have already won in an age when most conservatives understand and accept that gay marriage is here to stay.

Third, some of the variation is not heritable. It’s random. In fact around half of it within the population. Some people may just be liberal for stochastic reasons. You aren’t going to get rid of this with selection.

Perhaps most essential in terms of theory: frequency dependence. The dynamics of human interaction and decision making are such that the frequency of liberals declining might have an impact on their fitness. To give a weird example, perhaps an economically post-Malthusian society needs a certain number of sub-replacement liberals who engage in particularly productive work to maintain itself. If society slouched rapidly back toward Malthusianism perhaps everyone would just trudge along at replacement.

The big picture problem is assuming constant directional selection and exhaustion of heritable variation is all well and good when you are selecting for wax-seed oil, but human societies are non-linear systems which are subject to big shocks. They aren’t controlled agricultural genetic experiments.

Finally, let me use an analogous case to make an empirical objection. Many people tell me that the future will be religious due to the same dynamics above. This despite the century long trend toward secularization (parenthetical, God is Back was an ill-timed books, as the United States was shifting toward secularization at that time).

But I want to go back further. France was the first nation to start the demographic transition. In the early 19th century the secular elite was worried about the fertility of devout Roman Catholics, in particular the Poles who were arriving. The secular future they envisioned was threatened. It’s been nearly 200 years since these worries, and in those 200 years France has become more and more secular.

My point with this illustration is that if your theory can not predict the past, it can’t predict the future. At least not robustly. Liberal people will always be with us. So will shy people. And atheists too. They may wax and wane, but human variation persists. On the evolutionary genetic level I think frequency dependent dynamics are such that the fait, in the medium term, of low fitness traits is generally to become oddballs, not extinct. And once they are odd they may become fortunes favorites….

* For real, is that his real name?

Just because it’s not hereditary does not mean you can affect it

A comment below from Andrew:

Love to see a post about which human traits worth caring about are notable for having little or no hereditary component. It is all good and well to know what we cannot change, but it makes more sense to focus personally and as a parent on those things that aren’t genetically preordained.

This is a common sentiment I’ve seen. If you haven’t read The Nurture Assumption, please do so. I’d say a substantial reason I think that The Blank Slate is a good book is that Steven Pinker promoted Judith Rich Harris’ work.

With that out of the way: the implication in the comment above is that hereditary traits are the ones you have least control over, so you should focus on the non-hereditary traits. To some extent there is truth in this. Micronutrients are important. You don’t want to turn you children into cretins.

But a major problem with the idea that we can impact environmental impacts on characteristics is that on many traits we don’t know what those environmental impacts are. You can take a behavior genetic model and come to the following conclusion: within the population 50% of the variation is due to genes, 40% of the variation is due to non-shared environment, and 10% of the variation is due to shared environment (parents). We don’t really usually know what the non-shared environment means. It might be just developmental noise. It might be epistatic genetic effects. Or, in relation to behavior, it might be peer group, as Judith Rich Harris asserts.

We just don’t know. What that means is that the hereditary components are what you have legitimate effective control over through mate choice. And shared environment. These two combined are not nothing. And of course there is the impact of nation or community on the environmental in which propensities are expressed.

Addendum: The non-shared environmental variance was once explained to me as a “noise” factor. Just to give you a sense of how well we understand it.