Modeling genotype to phenotype: a search for "natural" objects of selection
Brian Cleary
Faculty of Computing and Data Sciences, Boston University
Primer: Flux Balance Analysis and shadow prices in constrained optimization
Aedan Brown
Ph.D. candidate, Boston University
Meeting: Modeling genotype to phenotype: a search for the right variables and the “natural” objects of selection
Simple genotype-to-phenotype maps allow for easy prediction about
selective pressure and evolution at the level of genotype. For most traits, the genotype-to-phenotype map is non-trivial, making straightforward predictions challenging. In this talk, we examine evolutionary dynamics with a non-trivial and mathematically tractable genotype-to-phenotype map: metabolism. Using evolutionary simulation and theory, we find that selection at the level of genes may be non-identifiable and at best paints an incoherent picture, whereas as selection on a preferred direction in phenotype space, the collective mode, appears to be a natural object of selection. We develop and analyze a theory of collective modes and help clarify the relative roles of genotype and phenotype in evolution, as well as providing an explanation for variability in polygenic traits.