192
J. D. Wall
the multiregional model (Wolpoff et al. 1984), which claims that modern humans
evolved from archaic humans simultaneously in Africa, Europe, and Asia (see Fig.
8.1 for a schematic of these models). Under the multiregional model, there would
not be strong systematic differences in levels of LD across continents, and we would
expect greater degrees of population differentiation (as measured by F ST ) between
different human continental groups.
One potential model for explaining global patterns of human genetic variation
is the serial bottleneck model (Ramachandran et al. 2005; DeGiorgio et al. 2009),
which posits that non-African populations experienced a series of founder bottlenecks as they left Africa and spread across the rest of the world. Under this model,
populations that are further away from the putative origin of modern humans (in
Eastern or Southern Africa) have experienced more bottlenecks and are expected to
have decreasing levels of variation and increasing levels of LD. Genomic data are
mostly consistent with these expectations (DeGiorgio et al. 2009; 1000 Genomes
Project Consortium 2010, 2012, 2015; Luca et al. 2011; Mallick et al. 2016), but
a denser sampling of the world’s populations is needed to more fully test this
theory. If the serial bottleneck model were mostly correct, it might help explain
why inferences of population history based on mtDNA are qualitatively similar to
ones based on autosomal data (see also Sect. 8.6).
8.5
Sex Ratio
Human effective population sizes vary not only across time but also between
males and females. In principle, this gender-specific difference can be estimated
by looking at patterns of genetic diversity on the X chromosome relative to the
autosomes. In a randomly mating population with equal numbers of males and
females, there are 3 X chromosomes for every four autosomes, and we expect to
observe a 3:4 ratio in levels of diversity in X vs. autosome comparisons. However,
if the numbers of breeding females and breeding males are unequal, there will be
a resulting skew in this 3:4 ratio (Caballero 1994). For example, in polygynous
societies, there is greater variation in male reproductive success, since many males
have no wives, while some have many. This in turn reduces the male effective
population size and leads to an increased ratio of X to autosome diversity.
However, there are many other factors that can affect relative levels of diversity
on sex chromosomes versus autosomes. Past changes in population size (e.g., population bottlenecks) can lead to temporary variation in the ratio of sex chromosome to
autosome diversity levels (Fay and Wu 1999), while recent positive selection, which
decreases levels of genetic variation due to hitchhiking effects, is expected to be
more effective on the X chromosome than the autosomes. (This is because recessive
advantageous mutations cannot easily spread on the autosomes but can on the X
chromosome since their advantageous effect is unmasked in males.) Analyses of
human polymorphism data show the effects of both of these evolutionary processes
(Hammer et al. 2008, 2010; Keinan et al. 2009; Arbiza et al. 2014). Overall, the ratio
of X to autosome diversity increases with increasing distance from genes, consistent
J. D. Wall
the multiregional model (Wolpoff et al. 1984), which claims that modern humans
evolved from archaic humans simultaneously in Africa, Europe, and Asia (see Fig.
8.1 for a schematic of these models). Under the multiregional model, there would
not be strong systematic differences in levels of LD across continents, and we would
expect greater degrees of population differentiation (as measured by F ST ) between
different human continental groups.
One potential model for explaining global patterns of human genetic variation
is the serial bottleneck model (Ramachandran et al. 2005; DeGiorgio et al. 2009),
which posits that non-African populations experienced a series of founder bottlenecks as they left Africa and spread across the rest of the world. Under this model,
populations that are further away from the putative origin of modern humans (in
Eastern or Southern Africa) have experienced more bottlenecks and are expected to
have decreasing levels of variation and increasing levels of LD. Genomic data are
mostly consistent with these expectations (DeGiorgio et al. 2009; 1000 Genomes
Project Consortium 2010, 2012, 2015; Luca et al. 2011; Mallick et al. 2016), but
a denser sampling of the world’s populations is needed to more fully test this
theory. If the serial bottleneck model were mostly correct, it might help explain
why inferences of population history based on mtDNA are qualitatively similar to
ones based on autosomal data (see also Sect. 8.6).
8.5
Sex Ratio
Human effective population sizes vary not only across time but also between
males and females. In principle, this gender-specific difference can be estimated
by looking at patterns of genetic diversity on the X chromosome relative to the
autosomes. In a randomly mating population with equal numbers of males and
females, there are 3 X chromosomes for every four autosomes, and we expect to
observe a 3:4 ratio in levels of diversity in X vs. autosome comparisons. However,
if the numbers of breeding females and breeding males are unequal, there will be
a resulting skew in this 3:4 ratio (Caballero 1994). For example, in polygynous
societies, there is greater variation in male reproductive success, since many males
have no wives, while some have many. This in turn reduces the male effective
population size and leads to an increased ratio of X to autosome diversity.
However, there are many other factors that can affect relative levels of diversity
on sex chromosomes versus autosomes. Past changes in population size (e.g., population bottlenecks) can lead to temporary variation in the ratio of sex chromosome to
autosome diversity levels (Fay and Wu 1999), while recent positive selection, which
decreases levels of genetic variation due to hitchhiking effects, is expected to be
more effective on the X chromosome than the autosomes. (This is because recessive
advantageous mutations cannot easily spread on the autosomes but can on the X
chromosome since their advantageous effect is unmasked in males.) Analyses of
human polymorphism data show the effects of both of these evolutionary processes
(Hammer et al. 2008, 2010; Keinan et al. 2009; Arbiza et al. 2014). Overall, the ratio
of X to autosome diversity increases with increasing distance from genes, consistent