210
L. S. Emery and J. M. Akey
diversity compared to other great apes (Kaessmann et al. 2001) and higher levels
of genetic diversity in African compared to non-African populations (Vigilant et al.
1991; Jorde et al. 2000; Tishkoff et al. 2009). Genetic data is consistent with several
models of dispersal, including a serial founder effect (Ramachandran et al. 2005;
Deshpande et al. 2009; Barbujani and Colonna 2010) or an isolation-by-distance
model (Handley et al. 2007). Regardless of the appropriate model, the essential
consequence of the peopling of different geographic regions is that each subsequent
dispersal was associated with a reduction in population size, often referred to as
“population bottlenecks” (Cavalli-Sforza 2007). Moreover, as humans dispersed
into new environments, selective pressure to adapt to emergent climates, diets,
and pathogens occurred, resulting in geographically restricted adaptation. In more
recent human history, particularly with the advent of agriculture, population sizes
expanded dramatically, which also likely created new selective pressures (CavalliSforza 2007). Indeed, it has been suggested that recent population growth has led to
rapid adaptation and an accelerated rate of evolution (Hawks et al. 2007).
The evolutionarily recent divergence of human populations and relatively high
rates of migration and admixture between them have the important consequence
that most of the variation observed in humans is found among all populations, and
not between them (Barbujani et al. 1997; Jorde et al. 2000), resulting in clinal
patterns of genetic diversity (Handley et al. 2007). Population history determines
which alleles are found where, thus limiting the available substrates for selection to
act upon. Indeed, the geographic distribution of SNPs that are highly differentiated
between populations (and possibly subject to selection) is virtually the same as the
distribution of randomly chosen SNPs used to determine population structure (Coop
et al. 2009). Such highly differentiated SNPs are important possible examples of
local adaptation, but other local adaptations are likely the result of subtle allele
frequency shifts and parallel adaptation in similar environments (Hancock et al.
2010a, b; Tennessen and Akey 2011). Accounting for the shared history between
populations is very important for identifying and interpreting the evidence for local
adaptions in specific populations (Tennessen and Akey 2011).
The population expansions and contractions experienced at various times in
human history also complicate the detection of advantageous alleles because it is
well-documented that such demographic events can produce patterns of variation
similar to those left by the hitchhiking effect and background selection (Tajima
1989; Przeworski 2002; Stajich and Hahn 2005; Li et al. 2012). Population
expansions can also mimic the genomic signatures of selection in certain conditions
(Excoffier et al. 2009). Selection events may also be associated with demographic
changes, such as population reduction caused by selective pressure, making the two
forces more difficult to distinguish (Li et al. 2012). It may seem that demography
and selection cannot be distinguished, but a reasonable way to address this problem
is to employ an outlier approach, which assumes that demographic events will
affect variation throughout the genome, while selective events act at individual loci
(Cavalli-Sforza 1966). Thus, advantageous alleles can be detected by identifying
loci that exhibit anomalous patterns of variation compared to the rest of the genome
L. S. Emery and J. M. Akey
diversity compared to other great apes (Kaessmann et al. 2001) and higher levels
of genetic diversity in African compared to non-African populations (Vigilant et al.
1991; Jorde et al. 2000; Tishkoff et al. 2009). Genetic data is consistent with several
models of dispersal, including a serial founder effect (Ramachandran et al. 2005;
Deshpande et al. 2009; Barbujani and Colonna 2010) or an isolation-by-distance
model (Handley et al. 2007). Regardless of the appropriate model, the essential
consequence of the peopling of different geographic regions is that each subsequent
dispersal was associated with a reduction in population size, often referred to as
“population bottlenecks” (Cavalli-Sforza 2007). Moreover, as humans dispersed
into new environments, selective pressure to adapt to emergent climates, diets,
and pathogens occurred, resulting in geographically restricted adaptation. In more
recent human history, particularly with the advent of agriculture, population sizes
expanded dramatically, which also likely created new selective pressures (CavalliSforza 2007). Indeed, it has been suggested that recent population growth has led to
rapid adaptation and an accelerated rate of evolution (Hawks et al. 2007).
The evolutionarily recent divergence of human populations and relatively high
rates of migration and admixture between them have the important consequence
that most of the variation observed in humans is found among all populations, and
not between them (Barbujani et al. 1997; Jorde et al. 2000), resulting in clinal
patterns of genetic diversity (Handley et al. 2007). Population history determines
which alleles are found where, thus limiting the available substrates for selection to
act upon. Indeed, the geographic distribution of SNPs that are highly differentiated
between populations (and possibly subject to selection) is virtually the same as the
distribution of randomly chosen SNPs used to determine population structure (Coop
et al. 2009). Such highly differentiated SNPs are important possible examples of
local adaptation, but other local adaptations are likely the result of subtle allele
frequency shifts and parallel adaptation in similar environments (Hancock et al.
2010a, b; Tennessen and Akey 2011). Accounting for the shared history between
populations is very important for identifying and interpreting the evidence for local
adaptions in specific populations (Tennessen and Akey 2011).
The population expansions and contractions experienced at various times in
human history also complicate the detection of advantageous alleles because it is
well-documented that such demographic events can produce patterns of variation
similar to those left by the hitchhiking effect and background selection (Tajima
1989; Przeworski 2002; Stajich and Hahn 2005; Li et al. 2012). Population
expansions can also mimic the genomic signatures of selection in certain conditions
(Excoffier et al. 2009). Selection events may also be associated with demographic
changes, such as population reduction caused by selective pressure, making the two
forces more difficult to distinguish (Li et al. 2012). It may seem that demography
and selection cannot be distinguished, but a reasonable way to address this problem
is to employ an outlier approach, which assumes that demographic events will
affect variation throughout the genome, while selective events act at individual loci
(Cavalli-Sforza 1966). Thus, advantageous alleles can be detected by identifying
loci that exhibit anomalous patterns of variation compared to the rest of the genome