226
L. S. Emery and J. M. Akey
models that lack the dynamic changes in selection strength that characterize
balancing selection.
On the other hand, negative selection acting on both populations will result in
lower differentiation between them at the site of selection. Both populations will
experience selective pressure to maintain the ancestral allele and have any arising
deleterious variants purged. Although specific examples of this phenomenon have
not been described, genes that show modest levels of silent human polymorphism,
but no divergence from chimpanzee at the protein level, are more likely to cause
Mendelian diseases (Barreiro et al. 2008). We can infer from this that sites under
pervasive purifying selection have limited capacity for population differentiation,
mainly because allele frequencies will be low at selected variants in both population;
however, linked neutral sites affected by background selection will actually show
increased levels of differentiation due to lower within-population diversity levels
(Charlesworth et al. 1997).
Positive selection acting on both populations can have several possible effects
on levels of differentiation at a locus (Fig. 9.3b–e). Although a non-negligible
proportion of identified targets of positive selection are shared among populations
at a local, regional, or global scale (Pickrell et al. 2009), selection may be acting
on vastly different causative alleles in different populations. Selection in two
given populations could be acting on two different functions of the same gene,
which would cause extremely high differentiation. Additionally, selection may be
driving two populations to different extremes of the same phenotype (Fig. 9.3b),
which would cause extremely high differentiation at the causative allele and linked
neutral sites (Charlesworth et al. 1997). Differentiation between two populations
under positive selection also depends on the variants available for selection to
act upon. Even if positive selection is moving two populations toward the same
trait, if different variants are available in each population, there will be high
levels of resulting differentiation (Fig. 9.3e). Several notable examples of this
kind of convergent evolution have been identified thus far. For instance, lactase
persistence alleles of LCT have arisen separately in European (Bersaglieri et al.
2004), Middle Eastern (Enattah et al. 2008), and East African (Tishkoff et al.
2007) populations. Each population possesses a different lactase persistence allele—
sometimes multiple alleles—with different haplotype backgrounds. So, although
positive selection is driving these populations toward the same phenotypic trait, the
result is still unusually high differentiation compared to the rest of the genome.
A recent genome-wide scan identified numerous SNPs exhibiting unusually high
levels of parallel divergence in two or more populations, suggesting convergent
evolution (Tennessen and Akey 2011). Migration rates between the populations will
also interact with positive selection to affect differentiation. If the beneficial allele
arises in one population and then spreads to the other through migration, there will
actually be a decrease in differentiation near the selected site, and the hitchhiking
region will be narrower (Fig. 9.3c) (Santiago and Caballero 2005). On the other
hand, if the beneficial allele is present in both populations before selection starts,
as in the case of selection from standing variation, differentiation will be higher
L. S. Emery and J. M. Akey
models that lack the dynamic changes in selection strength that characterize
balancing selection.
On the other hand, negative selection acting on both populations will result in
lower differentiation between them at the site of selection. Both populations will
experience selective pressure to maintain the ancestral allele and have any arising
deleterious variants purged. Although specific examples of this phenomenon have
not been described, genes that show modest levels of silent human polymorphism,
but no divergence from chimpanzee at the protein level, are more likely to cause
Mendelian diseases (Barreiro et al. 2008). We can infer from this that sites under
pervasive purifying selection have limited capacity for population differentiation,
mainly because allele frequencies will be low at selected variants in both population;
however, linked neutral sites affected by background selection will actually show
increased levels of differentiation due to lower within-population diversity levels
(Charlesworth et al. 1997).
Positive selection acting on both populations can have several possible effects
on levels of differentiation at a locus (Fig. 9.3b–e). Although a non-negligible
proportion of identified targets of positive selection are shared among populations
at a local, regional, or global scale (Pickrell et al. 2009), selection may be acting
on vastly different causative alleles in different populations. Selection in two
given populations could be acting on two different functions of the same gene,
which would cause extremely high differentiation. Additionally, selection may be
driving two populations to different extremes of the same phenotype (Fig. 9.3b),
which would cause extremely high differentiation at the causative allele and linked
neutral sites (Charlesworth et al. 1997). Differentiation between two populations
under positive selection also depends on the variants available for selection to
act upon. Even if positive selection is moving two populations toward the same
trait, if different variants are available in each population, there will be high
levels of resulting differentiation (Fig. 9.3e). Several notable examples of this
kind of convergent evolution have been identified thus far. For instance, lactase
persistence alleles of LCT have arisen separately in European (Bersaglieri et al.
2004), Middle Eastern (Enattah et al. 2008), and East African (Tishkoff et al.
2007) populations. Each population possesses a different lactase persistence allele—
sometimes multiple alleles—with different haplotype backgrounds. So, although
positive selection is driving these populations toward the same phenotypic trait, the
result is still unusually high differentiation compared to the rest of the genome.
A recent genome-wide scan identified numerous SNPs exhibiting unusually high
levels of parallel divergence in two or more populations, suggesting convergent
evolution (Tennessen and Akey 2011). Migration rates between the populations will
also interact with positive selection to affect differentiation. If the beneficial allele
arises in one population and then spreads to the other through migration, there will
actually be a decrease in differentiation near the selected site, and the hitchhiking
region will be narrower (Fig. 9.3c) (Santiago and Caballero 2005). On the other
hand, if the beneficial allele is present in both populations before selection starts,
as in the case of selection from standing variation, differentiation will be higher