polygenic adaptation

Racimo F, Berg JJ & Pickrell JK 2018 Detecting polygenic adaptation in admixture graphs. Genetics 208:1565-1584.

  • if a trait is polygenic, positive selection may instead occur by concerted shifts at many loci that all contribute to the variation in a trait
  • over short time scales, these shifts are expected to be small
  • see Barton and de Vladar (2009) for polygenic dynamics under longer time scales
  • none of the allele frequency changes need to be large on their own for the phenotypic change to be large
  • this process is called polygenic adaptation, and may underlie major evolutionary processes in recent human history
  • pleiotropy and phenotype definition
  • we are necessarily limited by the traits that have been defined and studied by others
  • the use of previously established definitions makes it difficult to understand exactly why these variants may have been under selection in the past
  • standardized schooling, and, consequently, the concept of "educational attainment," was only invented and implemented widely in the last few generations
  • it is obviously nonsensical to discuss its evolution over the past tens of thousands of years
  • it is likely that the set of variants for which we find evidence of selection was associated with some (unknown) phenotype(s) in the past
  • given that selection on these variants likely took place >5 KYA, it may be difficult or impossible to identify what these were
  • if a selective agent is cultural, but the culture has since changed, it may be impossible to determine what actually occurred
  • all these variants are also likely pleiotropic (Simons et al. 2017), which makes it even harder to determine which phenotypes were truly targeted by selection
  • the trait-associated variants that we have used only explain a fraction of the narrow-sense heritability of their respective traits
  • as we have only looked at variants that have high probability of association with a trait, this fraction is small in most cases
  • the heritability for "educational attainment" is estimated to be ∼40% and educational attainment itself is strongly determined by environmental factors
  • the SNPs we used in this study [themselves a subset of all SNPs tested in the original GWAS (Okbay et al. 2016)] explain only 1.05% of the total variance for this particular trait
  • our method bears a close relationship to a number of methods for inferring changes in the rate of phenotypic evolution on species phylogenies over macroevolutionary timescales
  • our use of the Normal model of drift as an approximation to the Wright-Fisher diffusion is closely analogous to the use of Brownian motion models in some phylogenetic methods
  • it may also be worth exploring the relationship between Ornstein-Uhlenbeck models for phenotypic evolution on phylogenies (Uyeda and Harmon 2014; Khabbazian et al. 2016) and the aforementioned hypothetical extension of our method to include stabilizing selection, as the two processes are closely related (Lande 1976; Simons et al. 2017)