This long persistence time of HLA lineages was explained by balancing selection [49, 52, 53].
In the case of mismatch distributions, selective sweep or purifying selection may mimic the effects of demographic expansion by generating unimodal shapes (due to an accumulation of low frequency alleles); on the contrary, balancing selection may mimic the effect of population contraction by generating multimodal profiles (due to the maintenance of intermediate frequency alleles).
may have enabled humans and chimps to retain multiple lines of defense that can be called on when a pathogen evolves new weapons.
It is assumed that at any given location there will be balancing selection on egg depth that reflects the costs and benefits of an egg being placed shallow or deep.
Put another way, there may be balancing selection for the same ovipositor length in all first generation bivoltine crickets, while diversifying selection among populations with differing season lengths is the norm for crickets producing overwintering eggs.
This implies that malaria earlier on wasn't a sufficiently serious cause of mortality to maintain the mutation through balancing selection.
If either mutation had occurred long ago and been maintained by balancing selection, DNA differences would have accumulated around different copies of each allele.
The cubic splines reveal that the nonlinearity was in fact balancing selection acting on total length.
Because the balancing selection on female total length results from opposing selection pressures on longevity and daily fecundity, our results may be interpreted as evidence of a cost of reproduction (Roff 1992; Stearns 1992).
One explanation is that balancing selection
is acting on the allozyme loci, whereas mtDNA haplotypes at the same time are selectively neutral and frequencies change through drift.
Under the neutral scheme, adaptive mutations (whether subject to balancing selection
or positive selection for fixation) are vanishingly rare, and the vast majority of molecular differences among species results from the chance fixation of neutral variants.
For the TIM model, B = 1, and there is no balancing selection
. For the SAS-CFF model, B [greater than] 1, and there is a balancing component to natural selection.