soft sweep

Fagny M, Patin E, Enard D, Barreiro LB, Quintana-Murci L & Laval G 2014 Exploring the occurrence of classic selective sweeps in humans using whole-genome sequencing data sets. Mol Biol Evol 31:1850-1868.

  • we found that background selection does not alter the power to detect selection following a hard sweep model
  • we evaluated the power of iHS and DIND to detect polygenic adaptation, which was modeled here as weak positive selection acting on many independent loci
  • this model of polygenic adaption has been proposed as an alternative model to rapid genetic adaptation, in light of the highly polygenic architecture of many traits in humans
  • neither DIND nor iHS detected a selection signal at low values of 2Ns
  • low 2Ns values lead to small shifts in the frequency of the selected alleles, as predicted under a model of polygenic adaptation acting through weak selection
  • our results support the notion that conventional methods have little power to detect signatures of polygenic adaptation
  • we performed simulations of positive selection on standing variation, that is, a neutral or mildly deleterious allele that is already segregating in the population at a frequency greater than 1/2N suddenly becomes positively selected and increases in frequency
  • the power of iHS and DIND was found to decrease with increasing initial frequency of the selected allele
  • high initial frequencies reduce the signature of the sweep around the selected site (Przeworski et al. 2005)
  • no signals of positive selection on standing variation were detected (data not shown) when simulations were performed with low values of 2Ns (2Ns < 10)
  • the frequency shifts of the selected alleles were, as for polygenic adaptation by weak selection, too small to be detected
  • iHS and DIND are insensitive to background selection and underpowered for the detection of polygenic adaptation or recent positive selection on standing variation when the selected allele has an initial frequency of 0.2 or above
  • the signals of positive selection detected by DIND and iHS in WGS data sets should reflect the effects of recent, strong positive selection targeting either a newly arisen allele (i.e., hard sweep stricto sensu) or standing mutations with a preselection frequency lower than 0.2 (nearly hard sweep)
  • our results must be seen in the context of recent debates as to the prevalence of hard sweeps in the human genome
  • two recent studies have suggested that classic selective sweeps have been relatively rare during human evolution (Hernandez et al. 2011; Granka et al. 2012)
  • most of these "sweeps" could be explained by the widespread action of background selection (Hernandez et al. 2011)
  • the two haplotype-based statistics used are robust to background selection and underpowered for the detection of positive selection events other than hard, or nearly hard, sweeps
  • our results should, therefore, highlight only the occurrence of recent, strong positive selection