soft sweep
Messer PW & Petrov DA 2013 Population genomics of rapid adaptation by soft selective sweeps. TREE 28:659-669.
- contemporary evolutionary biology is afflicted by an odd dichotomy
- experimental evidence suggests that adaptation via selective sweeps is often rapid, involving multiple adaptive mutations that rise in parallel at the same locus
- population genetic models typically assume mutation-limited scenarios and hard selective sweeps
- this discrepancy reflects the confusion of two different definitions of the effective population size
- adaptation is not limited by mutation in many species
- the key parameter determining whether adaptation is mutation-limited is Θ = 2Neμ
- mutation limitation and, consequently, hard sweeps, correspond to scenarios where Θ ≪ 1
- when Θ is on the order of 1 or larger, adaptation is not limited by mutation and sweeps become soft
- this holds true regardless of whether adaptation involves recurrent de novo mutations or multiple alleles from the standing genetic variation
- whether adaptation is generally mutation-limited in the evolution of a species does not depend on the harmonic mean of variance Ne estimated over long timescales
- the dynamics of adaptation should therefore be determined by how large variance Ne has been during its evolutionary history on average, which is described by the arithmetic mean
- this difference can explain the observation that soft sweeps are ubiquitous even in species where Ne values inferred from the levels of neutral diversity are low
- the possible prevalence of soft selective sweeps puts pressure on the field of population genetics to develop a more sophisticated understanding of the nonmutation-limited regime
- the distinction between de novo mutations and standing variation becomes blurred because every mutation at every site exists in the population most of the time
- populations should be able to explore the genotype space efficiently and not remain stranded on local fitness peaks for long periods of time
- complex, multistep adaptations can arise quickly, with intermediate steps not necessarily reaching high population frequencies
- genetic drift will be weak most of the time
- the patterns and levels of neutral polymorphisms should be primarily determined by the stochastic effects generated by recurrent selective sweeps at closely linked sites, the so-called 'genetic draft'