Development, GxE Interactions, and the Challenges of Causation

Consider the following scenario. We identify some behavioral trait. We determine that it has reasonably high heritability. GCTA (or other similar techniques) confirm that more similar genomes tend to be more similar... [ view full abstract ]

Consider the following scenario. We identify some behavioral trait. We determine that it has reasonably high heritability. GCTA (or other similar techniques) confirm that more similar genomes tend to be more similar phenotypically (though the “molecular” heritability estimate isn’t identical). GWAS studies find a number of more or less replicable ‘hits,’ that together explain a small fraction of the variance (much smaller than the heritability). Variation in the trait between populations is significant – often as large as or larger than the within-population individual variation addressed by the aforementioned studies. Finally, on reflection it seems likely that the trait in question is wrapped up with particular social practices; perhaps it is because of the kind of social practices we engage in that the trait exists as a trait, or perhaps our social practices make it a trait that we can sensibly name and measure. Now consider the question: what causes the heritable variation in the trait that we observe? Even in the case of variation within populations – the heritable variation that is associated with genetic differences, where GWAS studies find a number of SNPs, etc. – the question turns out to be more difficult to answer coherently than one might have guessed. In this talk, I’ll develop an example (based loosely on a thought experiment of James Flynn) that shows the ways in which a trait’s heritability may be produced by the kind of gene X environment interaction / correlation that makes identify something as a cause seem challenging. How common such interactions may be, and to what extent they are responsible for some of the patterns seen in behavior genetics research, will likely remain open questions for some time. But exploring the possible ways in which the developmental pathway taken can be influenced by genetic differences can reveal the ways in which environmental differences can be both driven by, and drivers of, phenotypic outcomes.