weak selection - senecio7の日記

Lion S, Jansen VAA & Day T 2011 Evolution in structured populations: beyond the kin versus group debate. TREE 26:193-201.

biological assumptions
the parallel between the MS and IF methodologies is sometimes obfuscated because many practitioners of the IF methodology in fact make two additional, interrelated simplifying assumptions as a routine part of the IF partitioning
in such cases, the IF partitioning not only yields conceptual insight, but also provides a means of making analytical progress
(i) fitness is linearly related to genotype
(ii) selection is weak
(i.e. mutations have small phenotypic effects)

the selection coefficient is often still too difficult to calculate analytically because the expected value of a neighbour's genotype will depend on the action of the mutant allele
if we further assume that selection is weak, however, then these expected genotypes can be calculated under the assumption that the mutant allele is neutral
if mutations have small effects, as typically assumed in IF methodology, fitness will be approximately linear in the genotype
hence, these two assumptions are distinct, but interrelated

much of IF theory typically assume mutations of small phenotypic effects
it has sometimes been argued that IF cannot deal with strong selection while the MS methodology can
however, partitioning of the selection coefficient between within-group and between-group components does not make its analytical calculation any simpler under strong selection
in other words, the MS methodology does not provide a superior characterization of genetic structure under strong selection
in fact most MS models also typically resort to a weak selection approximation or assume linear fitness to make analytical progress
the relevant question is not what approach is inherently better, but how to derive analytical approximations to characterize genetic structure under strong selection in an empirically meaningful way

is one interpretation better than the other?
neutral measures of genetic associations are central components of the IF methodology
for neutral alleles, the expected value of a neighbour's genotype corresponds exactly to the genealogical definition of relatedness and can be calculated for any type of population structure using standard population genetics techniques
under weak selection, the genealogical relatedness among individuals provides a sufficient description of the population's genetic structure to make evolutionary predictions

in the light of a recent article [9] that gives a misleading representation of current IF theory, we stress that extensions of the theory we have presented exist for fluctuating demography [32–34], non-linear genotype-phenotype maps [35] and non-pairwise interactions [36]

mutation and strong selection
when mutations are rare and have a small phenotypic effect (weak selection), genetic and demographic structuring tend to be decoupled
the evolution of reduced parasite virulence in structured populations can be understood as a form of altruism among genetically related parasites, and thereby falls under the scope of social evolution theory
these results typically rely on the assumption that mutation rates are low and that selection is weak