pleiotropy & compensatory evolution

Pavlicev M, Norgard EA, Fawcett GL & Cheverud JM 2011 Evolution of pleiotropy: epistatic interaction pattern supports a mechanistic model underlying variation in genotype–phenotype map. J Exp Zool B316:371-385.

  • understanding changes in the interaction structure, rather than changes in single gene products, is critical for understanding phenotypic variation
  • edgetics
  • epistatic interactions of loci with the rQTL will typically affect only one of the two traits, indicating that pathways are character specific and enabling evolutionary individuation of the characters
  • epistasis at rQTL exhibits a higher proportion of sign epistasis
  • on average 39% of the interactions involve sign change
  • only 12% do so for epistasis involving non-rQTL
  • compensatory mutations have previously been associated with pleiotropic loci
  • they can alleviate the deleterious pleiotropic (side-) effects of otherwise adaptive mutations
  • pleiotropic adaptive effect of deleterious mutations may be the reason for their persistence in a population
  • it is perhaps not surprising to find these effects predominantly in trait-specific interactions producing variation in pleiotropy
  • in diallelic crosses between inbred derivatives of the selected lines, it is plausible that the interaction partners form coadapted gene complexes within lines
  • in the case of pleiotropic genes, the coadapted complex may correspond to a pleiotropic mutation and the corresponding compensating trait-specific mutations
  • a surprisingly high proportion of rQTL do not have a detectable additive genetic effect
  • the mutations are stabilizing one trait while the other is evolving
  • plausible scenarios for sign epistasis include the possibility that mutations underlying rQTL are responsible for an interaction between molecules
  • e.g., between proteins or between DNA and a protein
  • these mutations do not segregate within the strain
  • when crossed, these complexes start segregating
  • such compensatory effects have been documented for cis-trans elements at higher divergence levels
  • given that two traits are under selection regimes that require independent variation, the existence of genetic covariance between these traits is disadvantageous because it causes a suboptimal correlated response
  • in the presence of sign epistasis at the pleiotropic locus, the genetic covariance contribution of that pleiotropic locus can change sign, if the interaction is trait specific