4 Types of Natural Selection and Tests of Selection
71
called genetic hitchhiking; as a directionally selected variant increases rapidly
to a high frequency, genetically linked neutral variants are also driven to high
frequencies. Balancing selection is the second form of selection that influences
nearby neutral variants. The third form of selection is called background selection.
Unlike directional selection and balancing selection that involve advantageous
mutations, background selection results from the removal of deleterious mutations.
We will first describe the effect of hitchhiking on linked neutral variation (part
A). Second, we will describe how balancing selection shapes neighboring neutral
diversity (part B). Finally, we will focus on background selection (part C).
(A) Hitchhiking
Genetic hitchhiking occurs when linked neutral variants are driven to higher
frequencies together with a positively selected, advantageous mutation that is on the
way to fixation. Hitchhiking leaves several signatures in neutral diversity that can be
used to detect and quantify positive selection. One of the most important signatures
is a local decrease of neutral diversity around the selected mutation (Maynard
Smith and Haigh 1974; Kaplan et al. 1989). As the selected mutation increases
in frequency, at neighboring neutral sites, linked alleles increase in frequency, and
unlinked alleles decrease in frequency. If they do not recombine with the selected
mutation, alleles unlinked to the selected mutation disappear from the population.
Following a hitchhiking event, the loss of neutral diversity is most severe near the
selected mutation and becomes gradually less severe as one gets further and further
away from the selected mutation (Fig. 4.1). This is because very few recombination
events occurred between the selected and the nearest neutral variants that could
rescue the latter from disappearing. As the distance from the selected mutation
increases, more and more recombination events occur, and as a result, more neutral
variants survive. The resulting pattern is a local dip in the level of diversity called
a selective sweep (Fig. 4.1). Diversity has been swept away during the process of
hitchhiking.
The loss of genetic variants is only one of several of the signatures left by
hitchhiking on patterns of diversity. Hitchhiking also affects both the site frequency
spectrum (SFS) and the structure of haplotypes in a population. The site frequency
spectrum is the distribution of the frequencies of variants. In an equilibrium
population, most neutral variants arose recently and segregate at low frequencies.
Most of these low-frequency variants ultimately get lost. Indeed, being at low
frequency means that under drift these variants are much more likely to be lost
than fixed, since it is far easier for them to reach a proportion of 0 than 100%.
Only a small minority of low-frequency variants make it to intermediate or high
frequencies, and the SFS is skewed toward low-frequency variants. In population
genetics, two types of SFS are used, the folded SFS and the unfolded SFS. The
unfolded SFS is the distribution of frequencies of derived alleles. The derived allele
at a particular position is the most recent (i.e. mutant) allele, as opposed to the
ancestral, preexisting allele. Determining which allele is derived and which allele
is ancestral is called polarization. It requires the use of a closely related species,
close enough that it is very likely that both ancestral allele and allele found in the
71
called genetic hitchhiking; as a directionally selected variant increases rapidly
to a high frequency, genetically linked neutral variants are also driven to high
frequencies. Balancing selection is the second form of selection that influences
nearby neutral variants. The third form of selection is called background selection.
Unlike directional selection and balancing selection that involve advantageous
mutations, background selection results from the removal of deleterious mutations.
We will first describe the effect of hitchhiking on linked neutral variation (part
A). Second, we will describe how balancing selection shapes neighboring neutral
diversity (part B). Finally, we will focus on background selection (part C).
(A) Hitchhiking
Genetic hitchhiking occurs when linked neutral variants are driven to higher
frequencies together with a positively selected, advantageous mutation that is on the
way to fixation. Hitchhiking leaves several signatures in neutral diversity that can be
used to detect and quantify positive selection. One of the most important signatures
is a local decrease of neutral diversity around the selected mutation (Maynard
Smith and Haigh 1974; Kaplan et al. 1989). As the selected mutation increases
in frequency, at neighboring neutral sites, linked alleles increase in frequency, and
unlinked alleles decrease in frequency. If they do not recombine with the selected
mutation, alleles unlinked to the selected mutation disappear from the population.
Following a hitchhiking event, the loss of neutral diversity is most severe near the
selected mutation and becomes gradually less severe as one gets further and further
away from the selected mutation (Fig. 4.1). This is because very few recombination
events occurred between the selected and the nearest neutral variants that could
rescue the latter from disappearing. As the distance from the selected mutation
increases, more and more recombination events occur, and as a result, more neutral
variants survive. The resulting pattern is a local dip in the level of diversity called
a selective sweep (Fig. 4.1). Diversity has been swept away during the process of
hitchhiking.
The loss of genetic variants is only one of several of the signatures left by
hitchhiking on patterns of diversity. Hitchhiking also affects both the site frequency
spectrum (SFS) and the structure of haplotypes in a population. The site frequency
spectrum is the distribution of the frequencies of variants. In an equilibrium
population, most neutral variants arose recently and segregate at low frequencies.
Most of these low-frequency variants ultimately get lost. Indeed, being at low
frequency means that under drift these variants are much more likely to be lost
than fixed, since it is far easier for them to reach a proportion of 0 than 100%.
Only a small minority of low-frequency variants make it to intermediate or high
frequencies, and the SFS is skewed toward low-frequency variants. In population
genetics, two types of SFS are used, the folded SFS and the unfolded SFS. The
unfolded SFS is the distribution of frequencies of derived alleles. The derived allele
at a particular position is the most recent (i.e. mutant) allele, as opposed to the
ancestral, preexisting allele. Determining which allele is derived and which allele
is ancestral is called polarization. It requires the use of a closely related species,
close enough that it is very likely that both ancestral allele and allele found in the