Thread by @molly_przew

A new thread about @arbelharpak and my paper on the evolution of group differences in changing environments, now out in @PLOSBiology: journals.plos.org/plosbiology/ar… 1/n

https://twitter.com/arbelharpak/status/1314041322911735810

Complex trait variation arises from both genetic and environmental effects and complex traits evolve under the influence of both. Yet as human population geneticists, we sometimes ignore environmental effects or implicitly assume that they are fixed. 2/n

We do so even though we know that for some of the best studied cases of potential human adaptations, such as height, there have also been massive secular shifts in environmental effects: e.g., elifesciences.org/articles/13410 3/n

Depending on the trait, salient environmental effects may include not just where someone is born or when, but also socioeconomic factors such as education level, occupation, housing conditions etc… 4/n

In our paper, we consider how complex trait genetics are expected to evolve when environmental effects are not fixed 5/n

Some implications are kind of obvious. For example, if we don’t make the assumption that environmental effects are fixed, we cannot expect mean trait values across groups to be predicted by mean polygenic scores or vice versa. 6/n

Nor can we readily interpret evidence for polygenic adaptation. Shifts in the fitness optimum (it’s suddenly better to be taller) and shifts in environmental effects (current nutrition tends to decrease average height) have equivalent effects on the loci that affect height— 7/n

—in both cases there will be selection on genetic variants to increase height. So there can be polygenic adaptation for a trait, and group differences in mean genetic effects, *even when the trait optimum has not changed*. 8/n

The idea that there could be lots of polygenic adaptation in the face of trait stasis is nothing new: e.g., “But this intense natural selection does not produce rapid evolution. On the contrary, it stabilizes the species.” Haldane 1956 9/n

But it has enduring relevance for comparisons of human polygenic scores across the world—it means that differences in genetic effects cannot be interpreted as differences in trait optima. Not without knowledge of the direction and magnitude of environmental effects. 10/n

Thanks to two PLoS Biology reviewers and numerous others for super helpful comments on earlier versions of this paper: