Evolut 3417

O R I G I NA L A RT I C L E
doi:10.1111/j.1558-5646.2010.01136.x
SEX LINKAGE, SEX-SPECIFIC SELECTION, AND

THE ROLE OF RECOMBINATION IN THE
EVOLUTION OF SEXUALLY DIMORPHIC
GENE EXPRESSION
Tim Connallon1,2 and Andrew G. Clark1
Downloaded from https://academic.oup.com/evolut/article/64/12/3417/6853464 by guest on 17 March 2024

1
Department of Molecular Biology & Genetics, Cornell University, Ithaca, New York 14853-2703
2
E-mail: tmc233@cornell.edu
Received March 21, 2010

Accepted September 10, 2010
Sex-biased genes—genes that are differentially expressed within males and females—are nonrandomly distributed across animal
genomes, with sex chromosomes and autosomes often carrying markedly different concentrations of male- and female-biased
genes. These linkage patterns are often gene- and lineage-dependent, differing between functional genetic categories and be-
tween species. Although sex-specific selection is often hypothesized to shape the evolution of sex-linked and autosomal gene
content, population genetics theory has yet to account for many of the gene- and lineage-specific idiosyncrasies emerging from
the empirical literature. With the goal of improving the connection between evolutionary theory and a rapidly growing body of
genome-wide empirical studies, we extend previous population genetics theory of sex-specific selection by developing and ana-
lyzing a biologically informed model that incorporates sex linkage, pleiotropy, recombination, and epistasis, factors that are likely
to vary between genes and between species. Our results demonstrate that sex-specific selection and sex-specific recombination
rates can generate, and are compatible with, the gene- and species-specific linkage patterns reported in the genomics literature.
The theory suggests that sexual selection may strongly influence the architectures of animal genomes, as well as the chromosomal
distribution of fixed substitutions underlying sexually dimorphic traits.
KEY WORDS: Antagonistic pleiotropy, epistasis, sex-biased genes, sex chromosomes, sexual antagonism.
Sexual dimorphism is common among animal species, and is From a population genetics perspective, the “sexes as
thought to reflect sexually discordant natural or sexual selection species” analogy is less apt. Excluding the few genes that are
(Darwin 1871; Andersson 1994). The idea that sexual dimorphism limited to a single sex (e.g., Y-linked genes in males; W-linked
reflects differential adaptation—that sex-specific selection drives genes in females), males and females carry the same set of genes
evolutionary divergence between the sexes—is clearly articulated and make equal reproductive contributions to each generation.
by Trivers (1972), who states: Males and females may be exposed to different patterns of natural
and sexual selection, yet “gene flow” between the sexes is unre-
One can, in effect, treat the sexes as if they were different
species, . . . female “species” usually differ from male species stricted and should constrain intersexual divergence. Although it
in that females compete among themselves for such resources is clear that sexual dimorphism does evolve, the genetic basis and
as food but not for members of the opposite sex, whereas sequence of evolutionary events that permit males and females to
males ultimately compete only for members of the opposite
diverge is not well known.
sex, all other forms of competition being important only in-
sofar as they affect this ultimate competition. (Trivers 1972, The evolution of sexual dimorphism has been conceptualized
p. 153) with two models (Darwin 1871; Fisher 1958; Rhen 2000; Coyne

C 2010 The Author(s). Evolution
C 2010 The Society for the Study of Evolution.
3417 Evolution 64-12: 3417–3442

T. C O N NA L L O N A N D A . G . C L A R K
et al. 2008). Under a “sex-limited” model, sexual dimorphism lations between different traits will also constrain adaptive evo-
evolves by selection of mutations with sex-limited phenotypic lution, yet it is unclear how strongly pleiotropy might hinder the
effects. If sex-limited mutations are available, the evolution of evolution of sexual dimorphism or whether such constraints might
sexual dimorphism is not problematic. For most mutations, how- differentially impact sex chromosomes and autosomes. Third, the
ever, at least some degree of expression is expected in both sexes current theory is almost entirely deterministic and focuses on
(see Lande 1980; Rice 1984; Rhen 2000; Morrow et al. 2008; conditions in which selection favors the invasion of a rare allele,
Poissant et al. 2010; though the magnitude may be sex-specific: rather than the probability and rate of sex-specific divergence,
e.g., Mackay 2001). Consequently, the evolution of sexual dimor- which is governed by selection, mutation, and genetic drift (Crow
phism might require genetic substitutions at multiple loci, with and Kimura 1970; Ohta 1992). Contrasts between chromosomes
genetic interactions underlying intersexual divergence. Such a should be based on the relative invasion probabilities of individual
multistep process could potentially involve sexually antagonistic mutations, and waiting times for the evolution of sexual dimor-
divergence—correlated evolution between the sexes that increases phism. Finally, the relationship between phenotype and fitness

adaptation of one sex and reduces it for the other—followed by the has not been consistently articulated by theory, in some cases
evolution of sex-limited expression of sexually antagonistic traits leading to opposing predictions about the relationship between
(Lande 1980; Rice 1984; Bonduriansky and Chenoweth 2009; van sex linkage and sexual antagonism (e.g., Rice 1984; Patten and
Doorn 2009; Stewart et al. 2010). Haig 2009; Fry 2010). Fortunately, there is a considerable body
Large proportions of animal genomes are differentially ex- of research linking gene expression variation to fitness variation
pressed between the sexes (Parisi et al. 2003; Ellegren and Parsch (see below), which can be used to refine sex-specific selection
2007), with “sex-biased” gene expression potentially decoupling models and incorporate biologically plausible parameterization.
male and female development and permitting adaptive sexual dif- With the goal of integrating empirical patterns of sex-biased
ferentiation (Williams and Carroll 2009). The genomic distribu- gene expression and population genetics theory, we developed and
tion (i.e., chromosomal linkage patterns) of sex-biased genes may analyzed a series of two-locus models for the evolution of sex-
also provide clues about the evolutionary processes and genetic biased expression in response to selection for gene expression
basis underlying sex-specific divergence (Mank 2009). Despite divergence. We separately consider three general evolutionary
substantial variability between species, there is an emerging con- routes toward sexually dimorphic gene expression: (1) the fixa-
sensus that sex-biased genes are nonrandomly distributed between tion of sex-limited and tissue-specific mutations that are uncon-
sex chromosomes and autosomes (e.g., Reinke et al. 2000; Wang strained by pleiotropy or sexual antagonism and directly generate
et al. 2001; Parisi et al. 2003; Ranz et al. 2003; Khil et al. 2004; sexual dimorphism; (2) the sequential fixation of sexually antago-
Kaiser and Ellegren 2006; Storchova and Divina 2006; Sturgill nistic or pleiotropically constrained alleles and alleles at modifier
et al. 2007; Mueller et al. 2008; Mank and Ellegren 2009; Meisel loci, which interact epistatically to generate sexual dimorphism or
et al. 2009; Mořkovský et al. 2010). Such patterns may suggest tissue-specific divergence; and (3) the co-invasion and simultane-
an important role for sexual antagonism during the evolution of ous fixation of linked alleles that are not individually favored by
sexual dimorphism (e.g., Parisi et al. 2003; Rogers et al. 2003; selection, but are beneficial in combination. For each scenario, we
Oliver and Parisi 2004; Connallon and Knowles 2005; Ellegren analytically characterize the necessary conditions and timescale
and Parsch 2007; Gurbich and Bachtrog 2008; Mank 2009; In- of sex-biased gene evolution under X-, Z- and autosomal link-
nocenti and Morrow 2010)—an interpretation inspired by popu- age. The theory provides a general framework for the evolution
lation genetics theory that predicts unique evolutionary fates for of sex-biased gene expression and generates a set of specific and
sex-linked versus autosomal sexually antagonistic mutations (mu- testable predictions.
tations beneficial to one sex and deleterious to the other; Pamilo
1979; Rice 1984; Patten and Haig 2009; Fry 2010).
Nevertheless, the connection between sexually dimorphic Model
gene expression patterns and population genetics theory is incom- To accommodate the potential impact of both linkage and epis-
plete for several reasons. First, although sexual antagonism is of- tasis, we first developed two-locus, biallelic population genetic
ten invoked, it does not by itself generate sexual dimorphism, and models with arbitrary fitness assigned to each genotype and sex
previous theory has mostly neglected epistasis between sexually (Table 1). Because we are interested in contrasting patterns of
antagonistic alleles and mutations that modify intersexual genetic evolution on sex chromosomes and autosomes, the loci are as-
correlations (but see Rice 1984). The role of epistasis is likely sumed to be both X-linked, both Z-linked, or both autosomal. We
to be profound: recent research reveals that interactions between discuss the potential role of interchromosomal interactions within
different cis-regulatory motifs often underlie sexually dimorphic the discussion. The distance between interacting loci is presented
expression (Williams and Carroll 2009). Second, genetic corre- as a function of the recombination rate between them. Because
3418 EVOLUTION DECEMBER 2010

S E X L I N K AG E A N D S E X UA L D I M O R P H I S M
Table 1. Genotypes and corresponding sex-specific fitness parameters.
Genotype: Diploid Loci: Autosomes, Z-linked Loci in Males, X-linked Loci in Females
A1 A1 B1 B1 A1 A2 B1 B1 A1 A1 B1 B2 A1 A2 B1 B2 A2 A2 B1 B1 A1 A1 B2 B2 A2 A2 B1 B2 A1 A2 B2 B2 A2 A2 B2 B2
Female fitness: f 11 f 21 f 12 f 22C , f 22R f 31 f 13 f 32 f 23 f 33

Male fitness: m11 m21 m12 m22C , m22R m31 m13 m32 m23 m33
Z-linked in Females X-linked in Males
Genotype: A1 B1 A2 B1 A1 B2 A2 B2 A1 B1 A2 B1 A1 B2 A2 B2
Female fitness: f1 f2 f3 f4 m1 m2 m3 m4
The haplotype configuration A1 B1 /A2 B2 corresponds to fitness f 22C and m22C ; haplotype configuration A1 B2 /A2 B1 corresponds to f 22R and m22R .

recombination is potentially sexually dimorphic, sex-specific re- tual: because trans-acting mutations are likely to influence many
combination rates between the loci are given by parameters rf more downstream targets than cis-mutations, disruptive effects
and rm for females and males, respectively. Recursions for the of trans-acting mutations are generally expected to be delete-
four possible haplotypes (A1 B1 , A2 B1 , A1 B2 , A2 B2 ) are developed rious and will play a relatively small role during evolutionary
in Appendix 1, and follow the sequence of (a) birth, (b) selection, divergence (Stern 2000; Prud’homme et al. 2007; Wray 2007).
(c) recombination, and (d) syngamy. Nevertheless, as shown below, the evolutionary trajectories of
Because the population mutation rate in animals—for exam- pleiotropic cis-regulatory mutations can differ between sex chro-
ple, 4Ne u per autosomal locus, per generation—is expected to be mosomes and autosomes, and these results provide insight into
less than one, we assume that each locus will initially be fixed the evolutionary dynamics of hypothetical trans-acting mutations
or nearly fixed for a common allele (Crow and Kimura 1970). with sex-specific fitness consequences. We return to this subject
Consequently, evolutionary outcomes of selection at single loci in the discussion.
or pairs of loci will be governed by patterns of selection acting on To parameterize the fitness effects of gene expression muta-
rare genetic variants. Under this assumption, the strength of selec- tions, we consider the following model for the expression-fitness
tion, and probability of invasion for rare alleles or haplotypes can landscape. We assume that there is an ancestral state where selec-
be addressed analytically with local linearized stability criteria. tion in males and females favors the same pattern of expression
A common genotype is susceptible to the evolutionary invasion and that the population has evolved to this gene expression op-
of a rare allele or haplotype when the leading eigenvalue for the timum, with gene expression maintained by stabilizing selection
system, that is the largest eigenvalue of the Jacobian matrix for near the fitness peak. Following a change in the environment
the set of recursions, is greater than one. Furthermore, the leading (ecological; social; or genetic), the fitness landscape may diverge
eigenvalue provides information about the probability of invasion between the sexes, thereby displacing one sex from its optimum
versus loss of the rare allele, as described further below. For each and potentially generating directional selection within that sex
system of inheritance—autosomal, X- and Z-linkage—there are toward the new fitness peak (Fig. 1; stabilizing selection persists
three potential leading eigenvalues, which are each presented in for the other sex).
Table 2. Details of the recursions, as well as the stability analysis, Prior theory and data suggest that gene expression-fitness
can be found in Appendices 1 and 2. landscapes are likely to be concave in the vicinity of their expres-
sion optimum (e.g., Wright 1934; Charlesworth 1979; Kacser and
Burns 1981; Dekel and Alon 2005; Lunzer et al. 2005; Phadnis
FITNESS AS A FUNCTION OF GENE EXPRESSION
and Fry 2005; Kalisky et al. 2007; Bedford and Hartl 2009; Zhang
Throughout the analysis, we focus on evolutionary divergence in
et al. 2009). There are two lines of evidence to suggest that fitness
gene expression due to cis-regulatory substitutions. This partic-
surfaces are generally concave:
ular focus is empirically justified by the disproportionate contri-
bution of cis-relative to trans-regulatory changes during species (1) The best supported theory for allelic dominance (Wright’s
divergence for gene expression and morphology (e.g., Wittkopp physiological theory: Wright 1934; metabolic control the-
et al. 2004; Wray 2007; Carroll 2008; Wittkopp et al. 2008; Graze ory: Kacser and Burns 1981; empirical support based on mu-
et al. 2009), as well as the important role of cis-regulatory in- tant fitness effects—for example, Charlesworth 1979; Orr
teractions during sex-biased gene regulation and morphological 1991; Phadnis and Fry 2005; and direct gene expression-
divergence between the sexes (Williams and Carroll 2009; also based assays—for example, Kacser and Burns 1981;
see Loehlin et al. 2010a,b). An additional justification is concep- Middleton and Kacser 1983; Hartl et al. 1985) is based
EVOLUTION DECEMBER 2010 3419

Table 2. Candidate leading eigenvalues for the two-locus models of sex-specific selection, with alleles A1 and B1 fixed.
Locus A Locus B Both loci (A and B interacting)

f (1−r )
Autosome Linkage: λA = f 21
+ 2m
2 f 11
m 21
λ B = 2ff1211 + 2m
m 12
λ AB = 22C2 f11 f + m 22C2m(1−r m)

11
11
11
f 22C (1−r f )
X-Linkage: λ A = 4ff2111 1 + 1 + 8 ff11 m2
m
λ B = 4
f 12
f
1 + 1 + 8 f 11 m 3
f m
λ AB = 4 f
1 + 1 + 8 f
f 11 m 4
m (1−r )

21 1 11 12 1 11 22C 1 f

Z-Linkage: λ A = 4m 11 1 + 1 + 8 m 21 f1
m 21 m 11 f 2
λ B = 4m 11 1 + 1 + 8 m 12 f1 λ AB = 4m 11
m 12 m 11 f 3 m 22C (1−rm )
1 + 1 + 8 m 22C f1 (1−rm )
m 11 f 4

Figure 1. Divergent expression-fitness landscapes generate conflicting selection over gene expression. The vertical gray lines represent
the wild-type gene expression at the time point when the fitness landscapes diverge between the sexes. Arrows indicate the net direction
of selection: stabilizing selection in the top panels, and selection for increased expression in the bottom panels. (A) Sex-specific selection
without pleiotropy: selection favors a different level of gene expression in males and females. (B) Sex-specific selection with pleiotropy. A
gene is expressed in multiple tissues or during multiple time periods during development. Expression-fitness functions overlap between
the sexes within some contexts (context 1), and are divergent in others (context 2). Note that although we display symmetrical fitness
landscapes (i.e., deviations above and below expression optima are equally costly), this need not necessarily be the case. The theory
presented here permits asymmetries within or between male and female fitness landscapes.
on the premise that fitness benefits of gene expression fol- and Alon 2005). Although the concavity assumption is ul-
low a saturating function. Because expression of a gene timately an issue that can only be resolved by additional
is expected to also carry one or more costs—that is, en- experimental work, the current evidence strongly suggests
ergetic and metabolic trade-offs (Kacser and Beeby 1984; that fitness is a concave function of a gene’s expression
Hurst and Randerson 2000; Wagner 2005, 2007), competi- level. We therefore adopt this concave relationship through-
tion for translation (Gout et al. 2010), interference between out our analysis, with the caveat that our predictions are
molecular pathways (Lion et al. 2004), or toxicity (Clark subject to revision following future empirical findings.
1991)—it is a mathematical necessity that any composite
fitness function that incorporates both benefits and costs of If we assume that the fitness topologies of males and females
gene expression will be concave within the vicinity of the diverge by small increments relative to time, and that the majority
gene expression optimum (see Supporting information). A of mutations have small effects on gene expression relative to
balance between benefits and costs is particularly relevant the landscape, then we can model the fitness effects of mutations
for the expression and evolution of sex-biased genes, which along a concave fitness surface. Figure 2 presents a conceptual re-
appear to be condition-dependent, implying a significant lationship between concave fitness surfaces and their relationship
cost of gene expression (Wyman et al. 2010). to selection and dominance parameters. Fitness (w) as a func-
(2) Direct experimental measurement of benefits, costs, and tion of gene expression (x) can be formally developed with the
total fitness as a function of the gene expression pheno- relationship:
type, supports a concave fitness surface (Dean et al. 1986;

Dykhuizen et al. 1987; Dean 1989; Papp et al. 2003; Dekel w(x) = exp −c |x̂ − x|k ,

w(A2 A2 ) = exp{−c(x̂ − x1 − 2x)2 }. The dominance coeffi-

cient of A2 over A1 is therefore:
1 − exp {cx(2x̂ − 2x1 − x)}

hb =
1 − exp {4cx(x̂ − x1 − x)}
2(x̂ − x1 − x) + x 1 x
≈ = + ,
4(x̂ − x1 − x) 2 4(x̂ − x1 − x)
which can hypothetically range between 0.5 < hb < 0.75 (in this
parameterization, the additive case has hb = 12 ). Extending the
theory for an arbitrary kth-order fitness function, dominance of
beneficial mutations will range between 0.5 < hb < (2k − 1)/2k .

Figure 2. Conceptual relationships between gene expression
variation and dominance along a concave fitness surface. The solid
Deleterious mutations
line represents a hypothetical expression-fitness function with the When the population resides at the fitness peak (i.e., the wild type
expression optimum occurring at the filled circle. Mutations alter genotype produces optimal gene expression), all mutations are
expression by x when heterozygous and 2x when homozy- deleterious because they cause gene expression to deviate from
gous. The fitness effects of heterozygous and homozygous mu- the optimum. Fitness of a deleterious mutation in heterozygous
tations depend upon the position along the fitness surface. For and homozygous state (respectively) follows the functions: w(A1
a population that is fixed for alleles expressing at level b (at the
A2 ) = exp {− c(x)2 }, and w(A2 A2 ) = exp {− c(2x)2 }. For this
optimum), mutations decrease fitness and the dominance coeffi-
second-order function, dominance of each deleterious mutation
cient for a deleterious mutation is hd = α 1 /(α 1 + α 2 ) < 1/2. For a
population at position a or c on the fitness surface, beneficial mu-
is hd ≈ 14 . For a generalized kth-order function, dominance of
tations approaching position b will have dominance coefficients deleterious mutations is hd ≈ 1/2k .
hb = α 2 /(α 1 + α 2 ) > 1/2.
Pleiotropy
Previous models generally assume that sexually antagonistic se-
where x̂ is the gene expression optimum that maximizes fitness, lection arises from opposing directional selection between the
and c is a positive constant (c 1). Note that this function has an sexes (e.g., Fig. 1A). An alternative possibility arises for genes
inflection point as expression deviates strongly from the optimum. that are expressed in multiple tissues or developmental stages.
Within the vicinity of the optimum, the relevant parameter space Mutations for such genes are likely to have pleiotropic fitness
that we consider here, the fitness landscape will be concave and effects (e.g., Mank et al. 2008), with gene expression altered in
will not differ substantially from a parabolic function. The fitness more than one context of expression. When selection favors gene
landscape is concave when k > 1. To simplify the presentation, we expression divergence within a single context, but mutations alter
present derivations using a second-order function (k = 2, which expression in multiple contexts, gene expression evolves upon a
is a Gaussian fitness function) and present general results for complex fitness landscape with multiple, context-specific peaks
arbitrary values of k > 1. (Fig. 1B).
Pleiotropy may often be associated with gene expression al-
Beneficial mutations tering mutations because most genes are expressed in multiple
Consider a wild-type allele A1 , which is fixed within a pop- tissues or contexts (e.g., Chintapalli et al. 2007), yet the popula-
ulation and which causes gene expression at a level x1 . Fol- tion genetic consequences of pleiotropy have yet to be explicitly
lowing a shift in the environment, x1 no longer resides at the analyzed within the theoretical context of sex-specific selection
fitness optimum such that x1 = x̂. A2 is a rare mutation that (Fitzpatrick 2004; Mank 2009). To incorporate pleiotropy into the
changes gene expression from x1 to x1 + x when heterozy- theory, we follow Curtsinger et al. (1994) and model total fitness
gous, and to x1 + 2x when homozygous. For simplicity, as- per genotype as a multiplicative function of fitness within each
sume that wild-type expression is below the optimum, caus- context of expression (e.g., tissue or developmental stage). Thus,
ing selection to favor alleles that increase expression (because for a genotype “g” expressed in n contexts or tissues, total fitness is
the model is symmetrical, the opposite case yields the same re- wg = i=1n
wi . The model is roughly equivalent to a linear fitness
sults). Assuming that the homozygous genotype does not over- model in which the fitness benefits and costs of a mutation are
shoot the expression optimum, such that x̂ ≥ x1 + 2x and small for each context (1 − wi 1). For example, a biallelic locus
x > 0, the fitness of the three genotypes will be: w(A1 A1 ) = with one allele (A1 ) favored in j contexts and the other favored in
exp{−c(x̂ − x1 )2 }, w(A1 A2 ) = exp{−c(x̂ − x1 − x)2 }, and one context (A2 ), will follow the fitness scheme: w(A1 A1 ) = 1 − s;

j j
w(A1 A2 ) = (1 − sh) i=1 (1 − ti h i ) ≈ 1 − sh − i=1 ti h i ; and (Weinreich and Chao 2005)—in our case, where initial invasion
j j
w(A2 A2 ) = i=1 (1 − ti ) ≈ 1 − i=1 ti , where s and h represent of a sexually antagonistic or antagonistic pleiotropic allele gener-
selection and dominance coefficients with respect to the tissue that ates selection for a modifier allele causing sex- or tissue-specific
is under directional selection, and t1 -tj and h1 -hj are the selection divergence. As with the single-step scenario, SSWM assumptions
and dominance coefficients with respect to the j tissues that are permit us to approximate the two-step fixation process by analyz-
under stabilizing selection. Note that this is probably the sim- ing the waiting time for each invasion event. The expected waiting
plest model for pleiotropy, and is limited to the case in which the time until both substitutions is equal to the sum of individual wait-
genotype is the unit of selection. ing times until invasion. These approaches are used widely in the
theoretical literature, and are accurate under SSWM conditions
PROBABILITY OF INVASION (e.g., Stephan 1996; Weinreich and Chao 2005; Kim 2007).
For a population with common alleles A1 and B1 fixed or nearly
fixed (thus, stability is assessed at the equilibrium [A1 ] = [B1 ] =

1), the leading eigenvalue for the system of recursions provides Results
information about the strength and direction of selection acting EVOLUTIONARY ROUTES TOWARD SEXUALLY
on rare alleles or haplotypes. The leading eigenvalue minus one DIMORPHIC GENE EXPRESSION
(λL − 1) provides the equivalent of a selection coefficient at the Sexually dimorphic gene expression can evolve by three basic
particular point at which the system is analyzed (e.g., for a rare evolutionary pathways. For mutations with sex-limited and tissue-
mutation; Otto and Bourguet 1999; Otto and Yong 2002). For specific effects on gene expression, sexual dimorphism can evolve
our purposes, leading eigenvalues can be used to approximate the by single, unconstrained genetic substitutions, provided that the
probability of invasion for new mutations. Such probabilities can mutation alters expression within the specific context where di-
potentially refer to individual alleles (A2 or B2 ), or of haplotypes vergence is favored. For mutations that are constrained by sexual
(A2 B2 ). For 1 λL − 1 1/Ne , where Ne is the effective popu- antagonism or pleiotropy, sex-specific adaptation may arise as
lation size, the probability of establishment versus loss of a new a consequence of multiple genetic substitutions involving muta-
mutation is: tions whose epistatic effects generate a sex- and/or tissue-specific
phenotypic response.
2(λ L − 1)
P= ≈ 2(λ L − 1) (1) The fixation of combinations of mutations, although more
λ2L
restrictive than the fixation of sex-limited alleles, can occur either
(i.e., by branching process; Haldane 1927; Otto and Day 2007). sequentially or simultaneously. If one of the mutations at the two
The approximation is used in subsequent results, and its accuracy loci is individually favored in the wild-type genetic background,
is verified by simulation (see Figs. S1 and S2). it can increase from low initial frequency. Formally, this occurs
when λA > 1 or λB > 1 (see Table 2). As one allele increases in
TEMPORAL DYNAMICS frequency, selection will favor alleles at the other locus that inter-
To describe the relative rate at which sexual dimorphism evolves act positively with the invading allele such that the combination of
under different modes of inheritance, we develop expressions to alleles improves net fitness across the sexes. Such two-step pro-
characterize the mean waiting time until sex-specific expression cess leads to the eventual fixation of positively interacting pairs
divergence. Under the assumption of strong selection and weak of alleles that optimize expression divergence between the sexes.
mutation (SSWM; Gillespie 1984, 1991), where the strength of If mutations at the two loci are individually disfavored by
selection acting on a rare allele or haplotype is much greater selection (i.e., λA < 1 and λB < 1), but are favored in combi-
than the reciprocal of the population size (1 λL − 1 1/Ne , nation, sexual dimorphism can potentially evolve if the positive
as described above), and the rate of mutation is much smaller epistatic interactions between mutations are strong relative to the
than 1/Ne (i.e., Ne u 1), the temporal dynamics of individual recombinational distance between the loci (Crow and Kimura
substitutions are dominated by the waiting time until invasion of 1965; Weinreich and Chao 2005). Formally, this will occur when
an allele or haplotype (the time required for an allele to arise λAB > 1 (Table 2). In terms of gene expression evolution, an allele
within the population and then escape stochastic loss by genetic combination might be favored because it decouples gene expres-
drift), and the transit time for each selective sweep is negligible. sion variation between the sexes or between different tissues,
For models involving single-step evolutionary transitions and thereby reduces evolutionary constraints imposed by sexu-
(e.g., the invasion of alleles or haplotypes with sex-limited and ally antagonistic selection or antagonistic pleiotropy. Both types
tissue-specific effects), we calculate the mean waiting time until of epistasis could arise from fitness interactions between muta-
the invasion of a derived allele or haplotype. This basic approach tions at sex- or tissue-specific and nonsex-specific cis-regulatory
can also be applied to sequential invasion models with epistasis elements (see Williams and Carroll 2009).

The opportunity for each of these evolutionary transitions can and

systematically differ between sex chromosomes and autosomes.
3
Differential accumulation of sex-linked and autosomal substitu- TXfix ( f ) = , (3c)
tions is a function of several factors: (1) whether males or females 4N X u X sh b
are selected to diverge, (2) whether alleles are sex-limited in ex-
where NX represents the X-linked effective size, and uX the X-
pression or expressed in both sexes, (3) whether mutations have
linked mutation rate per generation with respect to the allele.
pleiotropic effects or instead alter expression within single tissues,
These results represent reciprocals of the instantaneous rate of
and (4) whether the relative rate of recombination differs between
adaptive substitution, which are often used in theories of molecu-
sex chromosomes and autosomes. Below, we present these re-
lar evolution, including contrasts between sex chromosomes and
sults by contrasting X-linked and autosomal inheritance. These
autosomes (Charlesworth et al. 1987; Kirkpatrick and Hall 2004;
results also apply to Z versus autosome contrasts, which represent
Vicoso and Charlesworth 2009a).
a mirror image, with results reversed across sexes.
The specific waiting times until substitution on the X and

autosomes is sensitive to the effective mutation rate on each chro-
THE FIXATION OF SEX-LIMITED, TISSUE-SPECIFIC
mosome (the ratio NX uX /NA uA is expected to be 3/4, but may
MUTATIONS
vary as a result of demography, life-history, or breeding system;
If a mutation only alters gene expression within the tissue and
Charlesworth 2001, 2009; Ellegren 2007, 2009; Hedrick 2007).
sex where divergence is beneficial, there will be no constraint to
Nevertheless, the relative rate of fixation for male-beneficial mu-
its invasion and eventual fixation. With recurrent mutation to the
tations will always be more sensitive to X-linkage and domi-
favored allele, the substitution will eventually become fixed, with
nance. Under the null expectation, 4NX uX = 3NA uA , the rela-
the waiting time determined by the mutation rate and strength of
tive waiting time until female-beneficial substitution is the same
selection in favor of the mutation. For an autosomal allele, the
on the X and autosomes: TXfix (f )/TAfix = 1. The relative time
mean waiting time until fixation will be:
until male-beneficial substitution is TXfix (m)/TAfix = 2hb , where
the substitution rate is always more rapid on the autosomes un-
1
Tfix = + τ, (2) der the concave fitness landscape model (i.e., hb > 1/2 when
N A u A Pr( f i x)
k > 1). X-linkage facilitates female-beneficial substitution as
where NA is the autosomal effective size (NA ≈ 2Ne ), uA is the long as 4NX uX > 3NA uA , and male-beneficial substitution when
autosomal rate of mutation to the sex-limited allele, Pr(fix) is the 2NX uX > 3NA uA hb , with the latter requirement always more re-
probability that the allele eventually becomes fixed, and τ rep- strictive for hb > 1/2. These results are in agreement with several
resents the transit time of the allele—that is, the time it takes previous studies concerned with “faster-X” molecular evolution
for the allele to change from frequency 1/NA to frequency one. (i.e., Charlesworth et al. 1987; Kirkpatrick and Hall 2004; Vicoso
Given the simplifying assumption of strong selection/weak mu- and Charlesworth 2009a).
tation (SSWM: 1 sh 1/Ne u; Gillespie 1984, 1991), the
INVASION OF PLEIOTROPIC MUTATIONS, FOLLOWED
second term has little impact and can be ignored (e.g., Stephan
BY MODIFIERS OF PLEIOTROPY
1996; Gillespie 2000; Weinreich and Chao 2005; Kim 2007). Fur-
thermore, the probability of fixation for a sex-limited beneficial Adaptive divergence may be constrained if expression-altering
mutation is approximately equal to the heterozygous selection co- mutations also affect tissues evolving under stabilizing selection.
efficient in favor of that mutation (shb , the result for sex-limited Consider the evolution of a pleiotropic locus A, at which a derived
selection on the allele). Equation (2) can therefore be modified mutation (A2 ; A1 represents the wild type) is beneficial within
to: one tissue and deleterious within another. For simplicity, we as-
sume that mutations have sex-limited effects on gene expression;
1 the evolutionary genomic consequences of sexual antagonism are
TAfix = . (3a)
N A u A sh b considered separately. Table 3 presents the fitness parameteriza-
tion under a model of antagonistic pleiotropy, where coefficients
s and t are positive and small (0 < s, t 1), and the dominance
Under the same conditions, except with X-linked inheritance, coefficient, hd = 1/2k (0 < hd < 0.5 for k > 0).
the mean waiting times for male-limited and for female-limited Under autosomal linkage and sex-specific selection in either
beneficial substitutions (respectively) are sex, a derived allele A2 will invade when:
3 th d
TXfix (m) = (3b) s> , (4a)
2N X u X s 1 − hd

Table 3. Fitness parameterization under the antagonistic pleiotropy model.
Genotype
A1 , A1 A1 A1 A2 A2 , A2 A2
Female-specific fitness effects

Female fitness f 11 = 1 − s f 21 = (1 − hd s)(1 − hd t) f 31 = 1 − t
Male fitness (A) m11 = 1 m21 = 1 m31 = 1
Male fitness (X) m1 = 1 – m2 = 1
Male-specific fitness effects
Female fitness f 11 = 1 f 21 = 1 f 31 = 1
Male fitness (A) m11 = 1 − s m21 = (1 − hd s)(1 − hd t) m31 = 1 − t
Male fitness (X) m1 = 1 − s – m2 = 1 − t

will become fixed when: whereas A2 B1 haplotypes do not; for a complete parameteriza-
t(1 − h d ) tion of the two-locus system, see Table S1). As before, invasion
s> , (4b) criteria for the B2 allele can be addressed with local linearized
hd
stability criteria. Autosomal and X-linked Jacobian matrices were
and will remain polymorphic when:
recalculated for an arbitrary equilibrium frequency of A2 , and sta-
th d t(1 − h d ) bility was determined with the modifier allele absent ([B2 ] = 0;
<s< . (4c)
1 − hd hd Appendix 3). Under autosomal inheritance and tight linkage be-
(these results are in agreement with those of Curtsinger et al. tween A and B (rm , rf 1/2, as expected for two cis-regulatory
1994). If A2 remains polymorphic, it will converge to the equilib- loci for the same gene), the leading eigenvalue with respect to the
rium frequency: B locus is:
(1 − p̂) f 22C (1 − r f ) + p̂ f 32
s(1 − h d ) − th d λB ≈
p̂ ≈ . (4d) 2w̄ f
(s + t)(1 − 2h d )
(1 − p̂)m 22C (1 − rm ) + p̂m 32
These conditions of invasion, fixation, and polymorphism + , (6a)
2w̄m
are the same under female-specific selection at an X-linked locus.
Under male-limited selection and X-linkage, the conditions for where w̄ f and w̄m represent mean female and male fitness as a
invasion and eventual fixation are the same (s > t), and there in function of the frequency of A2 , with B1 fixed. For X-linked loci,
no parameter combination permitting a stable polymorphism. the leading eigenvalue is:
When A2 is favored, a single copy will invade the population (1 − p̂) f 22C (1 − r f ) + p̂ f 32
with probability ∼2(λA − 1), where λA is presented in the first λB ≈
4w̄ f
row of Table 2. Under SSWM conditions, the mean waiting time
w̄ m 4
× 1 + 1 + 8 w̄m [(1− p̂) f22Cf (1−r . (6b)
until A2 invades on an autosome will be: f )+ p̂ f 32 ]
1 An analysis of these eigenvalues indicates that, given invasion

Taut (A2 ) = . (5a)
2N A u A (λ A − 1) of A2 prior to the introduction of B2 , and if an A1 B2 haplotype is
The expected waiting time under X-linkage will be: not strongly deleterious, selection will generally favor the invasion
of the modifier allele. As the net rate of recombination between
1
TX (A2 ) = . (5b) the loci increases relative to the strength of selection, invasion of
2N X u X (λ X − 1)
a modifier mutation is not guaranteed unless it is neutral in an A1
The invasion of mutations with deleterious pleiotropic effects genetic background, or A2 is common within the population (see
can subsequently select for a modifier allele (B2 ) that limits ex- Appendix 4 for details). Given our focus on epistatic interactions
pression divergence to the appropriate tissue, thereby maximizing in cis-, we focus our analysis of the sequential invasion model
fitness with respect to gene expression across different tissues. As- under a scenario of relatively tight linkage. We revisit this issue
suming allelic interactions in cis-, an A2 B2 (coupling) haplotype within the discussion.
will cause tissue-specific expression divergence, and A2 B1 haplo- The waiting time until B2 invades (following invasion and
types will be associated with both beneficial and costly expression equilibrium of the A2 allele) is sensitive to the equilibrium fre-
divergence (i.e., A2 B2 haplotypes eliminate the pleiotropic cost, quency of A2 , the rate at which A2 B2 haplotypes are formed by

mutation and recombination, and the strength of selection in favor

of the A2 B2 haplotype, once created. Accounting for each of these
factors (see Appendix 5), the mean waiting time until B2 invades
will fall within the boundary:
N A p̂(1 − p̂)(rm + r f ) + 2
2N A p̂v A (λ B − 1)[N A (1 − p̂)(rm + r f ) + 2]
1
< Taut (B2 ) < (7a)
2N A p̂v A (λ B − 1)
under autosomal inheritance, and:

2N X p̂(1 − p̂)r f + 3
< TX (B2 )
2N X p̂v X (λ B − 1)[2N X (1 − p̂)r f + 3]
1
< (7b)
2N X p̂v X (λ B − 1)
under X-linked inheritance (where NA and NX refer to the autoso-

mal and X-linked effective size, vA and vX represent chromosome-
specific mutation rates at the B locus, and λB is characterized by
eqs. 6a or 6b). When A2 is rare and A1 B2 is neutral, the waiting
time approaches the term on the left. As the equilibrium frequency
of A2 and the deleterious effect of A1 B2 increase, the waiting time
approaches the term on the right (Appendix 5).
Permissive conditions for gene expression divergence, along
with the expected waiting times until A2 and B2 invasion, are
presented in Figure 3. These representative results show that X-
linkage will constrain male-specific evolutionary divergence, but
will not constrain female-specific divergence. As with the sex- and
tissue-limited model presented previously, X-linkage will actually
enhance the rate of female-specific adaptation as long as 4NX uX > Figure 3. Sequential invasion conditions and waiting times un-
3NA uA , and will be equal when 4NX uX = 3NA uA . til tissue-specific gene expression divergence under antagonistic
These patterns are a natural consequence of evolution along pleiotropy. Pleiotropically expressed alleles (A1 and A2 ) follow the
a concave fitness surface. Concave fitness landscapes generate fitness parameterization from Table 3. Opportunities for invasion
dominance reversals: deleterious fitness effects act recessively, are described by the shaded bar at the base of each panel. The
curves represent the mean time until derived alleles invade at the
whereas beneficial effects are dominant (see above; Fig. 2). In fe-
pleiotropic and modifier loci, with the uniform curve representing
males, which are diploid throughout the genome, the expression of
the X-linked scenario and the circle-bearing curve representing
beneficial and deleterious fitness effects of individual mutations autosomal linkage. Results were obtained using the SSWM ap-
does not differ between the X and autosomes. However, female- proximations presented in the text. Results are shown for k = 2
limited selection can be more effective on the X because each (hd = 14 ), t = 0.01, and use the permissive left hand term of equa-
X-linked locus evolves within a female genome with two-thirds tions (7a and b); mutation rates per locus and per sex are u = v =
probability, in contrast to one-half probability for autosomes. In 10−8 ; X and autosome effective sizes are NA = 106 = 4NX /3. Results
males, benefits of expression divergence are marginally enhanced use dimorphic recombination parameters rf = 0.001 and rm = 0,
yet sexually monomorphic recombination yields a nearly identical
by X-linked inheritance because they are dominant, whereas costs
pattern.
of expression divergence are strongly enhanced by X-linkage be-
cause they act recessively. For male-specific adaptive divergence, INVASION OF SEXUALLY ANTAGONISTIC
pleiotropy should generate a strong bias off of the X because MUTATIONS, FOLLOWED BY MODIFIERS OF
pleiotropic constraints in males are exacerbated by hemizygous SEX-LIMITATION
expression. The severity of the bias will increase with the curva- Mutations that are expressed in both sexes can give rise to sexual
ture of the fitness landscape (i.e., as k increases). antagonism if they are favored in one sex but disfavored in the

Table 4. Fitness parameterization under the sexual antagonism model.
Genotype
A1 , A1 A1 A1 A2 A2 , A2 A2
Selection for expression divergence in females

Female fitness f 11 = 1 − sf f 21 = 1 − hd sf f 31 = 1
Male fitness (A) m11 = 1 m21 = 1 − hd sm m31 = 1 − sm
Male fitness (X) m1 = 1 – m2 = 1 − sm
Selection for expression divergence in males
Female fitness f 11 = 1 f 21 = 1 − hd sf f 31 = 1 − sf
Male fitness (A) m11 = 1 − sm m21 = 1 − hd sm m31 = 1
Male fitness (X) m1 = 1 − sm – m2 = 1

other. The invasion of each sexually antagonistic allele can subse- These alleles will go to fixation when
quently generate selection for alleles at secondary loci that limit 2s f (1 − h d )
expression divergence to the sex exposed to directional selec- sm > . (9b)
1 − s f hd
tion. Because invasion at these secondary loci are limited by prior
invasion at the sexually antagonistic locus, the opportunity for se- An X-linked, female-beneficial allele can invade when
quential invasion depends on whether each sexually antagonistic sm
sf > . (9c)
allele can invade and trigger the two-step process of divergence. 2 − h d (2 − sm )
The specific conditions for invasion have been described in sev- and will become fixed when:
eral previous studies (e.g., Kidwell et al. 1977; Pamilo 1979; Rice sm
1984; Albert and Otto 2005; Patten and Haig 2009; Fry 2010). sf > . (9d)
h d (2 − sm )
We summarize these results below, with relevant eigenvalues
X-linked polymorphism will be maintained under the condition
(Table 2) evaluated using the fitness landscape model described
above and in Table 4. 2s f h d 2s f (1 − h d )
< sm < (9e)
When divergence is favored in males, an autosomal male- 1 + s f hd 1 − s f hd
beneficial allele can invade when with the male- and female-beneficial alleles converging to the
s f hd respective equilibrium frequencies
sm > . (8a)
1 − h d (1 − s f ) sm − 2s f h d
p̂m ≈ (9f)
Such an allele will go to fixation when 2s f (1 − 2h d )
s f (1 − h d ) and
sm > . (8b)
h d (1 − s f ) 2s f (1 − h d ) − sm
p̂ f ≈ . (9g)
2s f (1 − 2h d )
The system will remain polymorphic when
The equilibria each require that hd < 0.5, as expected for the
s f hd s f (1 − h d )
< sm < . (8c) concave fitness landscape model.
1 − h d (1 − s f ) h d (1 − s f )
The expected time until invasion of a derived, sexually an-
Given our constraint that hd < 0.5, the male-beneficial polymor- tagonistic allele (A2 ) is given by equations. (5a) and (5b), with
phism will converge to the equilibrium frequency the relevant eigenvalues (λA ) parameterized using fitness values
sm (1 − h d ) − s f h d from Table 4. Coupling haplotypes with derived sexually antago-
p̂ ≈ . (8d) nistic and modifier alleles (A2 B2 ) will cause sex-specific expres-
(sm + s f )(1 − 2h d )
sion divergence, whereas repulsion haplotypes (A2 B1 ) generate
When females are selected to diverge, the same results apply, with expression divergence in both sexes (for a complete parameter-
sm and sf substituted. ization of the two-locus system, see Table S2). As with antago-
The invasion of an X-linked, male-beneficial mutation can nistic pleiotropy (again, assuming that A1 B2 haplotypes are not
occur when strongly deleterious), most conditions favoring invasion of a sex-
2s f h d ually antagonistic allele will also subsequently favor invasion of a
sm > . (9a)
1 + s f hd linked, cis-regulatory modifier of sex-limitation (see Appendix 4).

The general stability conditions and waiting times until the B2 al- m22C = 1 − hd sm , and m4 = 1; under female-specific selection,
lele invades (conditional on A2 reaching an equilibrium frequency fitness is defined as f 11 = 1 − sf , and f 22C = 1 − hd sf ; these are
greater than zero) are given by equations (6a–7b); see Appendix used to evaluate λAB ), simultaneous invasion of a male-beneficial
3 and 4 for the additional case of relatively loose linkage. haplotype will be favored on the autosomes when
Results for the sexually antagonistic modifier model show rm + r f
that gene expression divergence, involving the sequential inva- sm > . (10a)
1 + r f − h d (1 − rm )
sion of sexually antagonistic and modifier alleles, will generally
Simultaneous invasion of an autosomal, female-beneficial haplo-
be constrained by X-linked inheritance ( Figure 4). The conditions
type can occur when
permitting such a process are broader on the autosomes. Further-
more, the mean waiting time until the resolution of sexual antag- rm + r f
sf > . (10b)
onism (i.e., the invasion and fixation of A2 and B2 alleles) will be 1 + rm − h d (1 − r f )
abbreviated on the autosomes relative to the X. These results apply Under X-linkage, simultaneous invasion of a male-beneficial

to both male- and female-specific expression divergence. The un- haplotype can occur when
derlying mechanism for the bias off the X is similar to the case of
2r f
male-specific divergence under antagonistic pleiotropy: sexually sm > . (10c)
1+rf
antagonistic fitness costs (to the sex exposed to stabilizing selec-
tion) are exacerbated by X-linkage and minimized by autosomal Invasion of a female-beneficial haplotype is favored when
linkage, which limits opportunities for the invasion of sexually an- rf
sf > . (10d)
tagonistic alleles on the X. These results are in agreement with a 1 − h d (1 − r f )
recent analysis by Fry (2010), which focused on opportunities for
The waiting time until simultaneous invasion of a pair of
stable polymorphism on the X and autosomes, and runs counter
linked alleles depends on both the probability of invasion of an
to the intuition that sexual antagonism will generally lead to an
A2 B2 haplotype, as well as the rate at which A2 B2 haplotypes are
enrichment of sex-linked, sex-biased genes.
created by mutation or recombination. Assuming 1 λAB − 1
1/Ne , the probability of invasion and eventual fixation of a male-
SIMULTANEOUS INVASION OF EPISTATICALLY
beneficial haplotype will be
BENEFICIAL MUTATIONS
Strong selective constraints can limit opportunities for gene ex- Pr(A2 B2 |m, aut.) ≈ 2(λ AB − 1)
pression divergence. If individual mutations evolve under a net (1 − h d sm )(1 − rm ) − (1 − sm )(1 + r f )
=
purifying selection (averaged across tissues or across the sexes; (1 − sm )
formally, when λA < 1 and λB < 1; Table 2), and purifying (11a)
selection is strong enough to prevent the fixation of individual on the autosomes, and:
mutations by genetic drift (1 − λA , 1 − λB 1/Ne ), then sex-
specific evolutionary divergence requires the simultaneous sub- 1−rf 1
Pr(A2 B2 |m, X) ≈ 1+ 1+8 −2
stitution of positively interacting pairs of mutations. These pairs 2 (1 − sm )(1 − r f )
(11b)
of mutations can simultaneously invade when selection in favor of
a double-mutant haplotype is strong relative to the probability of on the X. The invasion probability of a female-beneficial haplo-
recombination between the mutations (Crow and Kimura 1965; type will be
Weinreich and Chao 2005; Kim 2007). (1 − h d s f )(1 − r f ) − (1 − s f )(1 + rm )
The specific parameter conditions that are conducive to si- Pr(A2 B2 | f, aut.) ≈
(1 − s f )
multaneous invasion can systematically differ between sex chro- (11c)
mosomes and autosomes, with invasion conditions for the derived on the autosomes, and
haplotype depending on: (1) whether the haplotype is favored in
(1 − h d s f )(1 − r f )
males or in females; (2) whether it is X-linked or autosomal; Pr(A2 B2 | f, X) ≈
2(1 − s f )
and (3) the degree of linkage (1 − rf , 1 − rm ) between the two
loci. The eigenvalue λAB (Table 2) characterizes whether an A2 B2 (1 − s f )
× 1+ 1+8 −2
haplotype can invade a population fixed for A1 B1 . Assuming that (1 − h d s f )(1 − r f )
each A2 B2 haplotype causes expression divergence within the ap- (11d)
propriate sex and tissue (i.e., sexually antagonistic or pleiotropic on the X.
fitness costs are absent in the beneficial haplotype A2 B2 ; under To calculate the rate at which A2 B2 haplotypes are generated,
male-specific selection, fitness is defined as m11 = m1 = 1 − sm , we assume (following previous theory; see Weinreich and Chao


Figure 4. Sequential invasion conditions and waiting times until sex-specific gene expression divergence under sexual antagonism.
Opportunities for invasion are described by the shaded bar at the base of each panel. The curves represent the mean time until derived
alleles invade at the sexually antagonistic and modifier loci, with the uniform curve representing the X-linked scenario and the circle-
bearing curve representing autosomal linkage. Results were obtained using the SSWM approximations presented in the text. Sexually
antagonistic alleles (A1 and A2 ) follow the fitness parameterization from Table 4. Results are shown for k = 2 (hd = 14 ), sf = 0.01 for the
case of directional selection in males, and sm = 0.01 for directional selection in females; mutation rates per locus and per sex are u = v =
10−8 ; X and autosome effective sizes are NA = 106 = 4NX /3. Results use dimorphic recombination parameters rf = 0.001 and rm = 0, yet
sexually monomorphic recombination yields a nearly identical pattern.
2005; Kim 2007) that the rate of recombination between repulsion types are roughly the same between the X and autosomes, whereas
haplotypes, A2 B1 and A1 B2 , is negligible (as expected when rm , male-beneficial haplotypes are constrained by X-linkage (Fig. 5).
rf 1, a requirement for invasion, and when A2 B1 and A1 B2 are In species with recombination in both sexes (e.g., mammals),
rare, as expected under purifying selection), and that parameters X-linkage expands the invasion conditions of both female- and
of mutation and purifying selection are the same at locus A and B. male-beneficial haplotypes (Fig. 6). These species-specific re-
Given these simplifications, the rate at which A2 B2 haplotypes are sults reflect an underlying constraint that will often prevent pairs
created is 2u2 /ω, where u is the mutation rate at each locus, per of mutations from invading simultaneously: rare, beneficial allelic
generation, and ω is the net strength of purifying selection acting combinations can spread if they tend to be co-inherited. Recombi-
on each deleterious haplotype. Incorporating the rate of creation nation between loci breaks apart beneficial genetic combinations,
and probability of fixation for A2 B2 , the mean waiting time until thereby preventing the invasion of adaptive genetic complexes. A
invasion will be closer look at conditions (10a–10d) provides an explanation for
the effect of recombination on the relative haplotype invasion con-
ωj
T (A2 B2 |i, j) = , (12) ditions on the X and autosomes. When there is no recombination
2N j u2 Pr(A2 B2 |i, j)
in males (rm = 0), the minimum selection coefficient required for
where j can refer to X or autosomal linkage, and i refers to male a female-beneficial haplotype to invade is identical between the
or female selection (subscripts m and f ). Characterizing the net X and autosomes; with male-specific selection, autosomal link-
strength of purifying selection on deleterious haplotypes requires age is more conducive to invasion when hd < 0.5, approximately
a detailed specification of the nature of sex-specific selection (assuming that rf 1). As rm increases, recombination becomes
acting on A2 B1 and A1 B2 haplotypes. Although space limitation effectively higher on the autosomes, the minimum conditions nec-
precludes a detailed analysis of all possibilities for the X and essary for invasion will increase for autosomal haplotypes, and
autosomes, we present waiting time results under different ratios: X-linkage will facilitate the invasion of both male- and female-
Rω = ωX /ωA . This is the most salient factor affecting the relative beneficial allelic combinations.
rate of A2 B2 creation on each chromosome. The waiting time until beneficial haplotypes become fixed is
Invasion conditions and waiting times, under the simulta- sensitive to the strength of positive selection, the effective rate of
neous substitution model, are presented in Figures 5 and 6. recombination, and the pattern of net purifying selection against
In species where recombination only occurs in females (e.g., individual derived alleles ( Figs. 5 and 6). When the strength of
Drosophila), the invasion conditions of female-beneficial haplo- purifying selection against disfavored alleles is greater on the X


Figure 5. Simultaneous invasion conditions and waiting times for epistatically beneficial haplotypes when males do not recombine.
Opportunities for invasion are described by the shaded bar at the base of each panel. The curves represent the mean time until double-
mutant haplotypes invade, with the uniform curve representing the X-linked scenario and the circle-bearing curve representing autosomal
linkage. The net autosomal purifying selection against A2 B1 or A1 B2 haplotypes is ωA = 0.001, with results for two ωX /ωA ratios shown.
Results are shown for k = 2 (hd = 14 ), sf = 0.01 for the case of directional selection in males, and sm = 0.01 for directional selection in
females; mutation rates per locus and per sex are u = v = 10−7 ; X and autosome effective sizes are NA = 106 = 4NX /3. Results were
obtained using the SSWM approximations presented in the text.
(i.e., ωX /ωA > 1), the underlying conditions permitting invasion binations for each gene, then epistatic transitions toward sexually
will be unaffected. However, the waiting time until invasion will dimorphic expression may be common.
increase on the X, due to a decrease in the rate at which A2 B2
haplotypes arise within the population.
Because our model invokes an interaction between two pre- Discussion
specified loci, and the ancestral population (A1 B1 fixed) is initially The ubiquity of sexual reproduction (Bell 1982; Rice 2002), along
two mutation steps away from a beneficial haplotype, the wait- with widespread observations of selection differences between
ing time for simultaneous invasion can be relatively long. Nev- males and females (Andersson 1994; Arqvist and Rowe 2005),
ertheless, the actual rate of epistatic coevolution might still be has inspired a large and growing body of evolutionary theory. Sex-
substantial. Simultaneous invasion is limited by the rate at which specific selection has implications for several important topics in
beneficial haplotypes arise within a population. This, in turn, de- evolutionary biology, including the maintenance of genetic vari-
pends on two factors: (1) the number of genes, at any given time, ation for fitness, the genomic architecture of species differences
that are exposed to sex-specific selection for expression diver- and sex-specific traits, the opportunity for and rate of adapta-
gence; and (2) the number of epistatically interacting mutational tion in sexually reproducing populations, and the evolution of
combinations that influence sex-specific expression variation at female mating biases (e.g., Manning 1984; Kondrashov 1988;
each gene. If there are a large number of genes under sex-specific Koeslag and Koeslag 1994; Whitlock 2000; Agrawal 2001; Rice
disruptive selection, or many epistatically beneficial allele com- and Chippindale 2001; Siller 2001; Lorch et al. 2003; Fedorka


Figure 6. Simultaneous invasion conditions and waiting times for epistatically beneficial haplotypes when both sexes recombine.
Opportunities for invasion are described by the shaded bar at the base of each panel. The curves represent the mean time until
double-mutant haplotypes invade, with the uniform curve representing the X-linked scenario and the circle-bearing curve representing
autosomal linkage. The results follow the same parameterization as presented in Figure 5, with rm = rf , and were obtained using the
SSWM approximations presented in the text.
and Mousseau 2004; Albert and Otto 2005; Pischedda and ing the evolution of sex-biased gene expression, by: (1) outlining
Chippindale 2006; Scotti and Delph 2006; Hadany and Beker the predictions of our models within the context of previous the-
2007; Calsbeek and Bonneaud 2008; Candolin and Heuschele ory; and (2) contrasting observed genomic patterns of sex-biased
2008; Bonduriansky and Chenoweth 2009; Cox and Calsbeek gene expression with theoretical predictions.
2009; Van Doorn 2009; Whitlock and Agrawal 2009; Blackburn
et al. 2010; Connallon 2010; Connallon et al. 2010; Cox and SEX-SPECIFIC SELECTION AND THE EVOLUTION
Calsbeek 2010). OF SEX-BIASED GENE EXPRESSION
The theory presented here builds upon several independent Sex-biased gene expression might evolve through adaptive shifts
contributions, including the population genetics of sexual antago- by males or by females. For example, a male-biased gene might
nism (e.g., Owen 1953; Haldane 1962; Kidwell et al. 1977; Pamilo evolve in response to stabilizing selection in males and selection
1979; Patten and Haig 2009; Fry 2010; and especially Rice 1984) for decreased expression in females, or through stabilizing selec-
and antagonistic pleiotropy (e.g., Curtsinger et al. 1994; Prout tion in females and selection for increased expression in males.
1999), X versus autosome molecular evolution (Charlesworth In other words, the observation of sex-biased expression is con-
et al. 1987; Kirkpatrick and Hall 2004; Vicoso and Charlesworth sistent with selection for either male or for female gene expres-
2009a), the evolution of peak shifts (e.g., Crow and Kimura sion divergence, and additional information is required to equate
1965; Weinreich and Chao 2005; Kim 2007), and the evolutionary sex-biased expression with male- or female-specific adaptive di-
and physiological basis of allelic dominance (e.g., Wright 1934; vergence (Connallon and Knowles 2005). To interpret genomic
Kacser and Burns 1981; Dekel and Alon 2005; Gout et al. 2010). patterns of sex-biased expression within the theoretical frame-
Below, we discuss the potential role of sex-specific selection dur- work developed here, the frequency of each evolutionary route

toward male- and female-biased gene expression must be known. Under a model of sequential coevolution between sexually
Within Drosophila, the answer appears to be clear: the evolu- antagonistic alleles and modifiers of sex-limited expression,
tion of male-biased gene expression typically involves expression sex chromosomes are less hospitable to both male- and
increases in males (male divergence); female-biased gene expres- female-biased genes. This result contradicts the widespread
sion involves expression increases in females (female divergence; intuition that sex linkage should generally promote the evo-
Vicoso and Charlesworth 2009b). Although we acknowledge that lution of sex-biased expression from an initially sexually
additional research will be required to assess the generality of antagonistic state—an expectation based on the assumption
this pattern, the following discussion is based on this clear pattern of constant allelic dominance for beneficial and deleteri-
from Drosophila. That is, our interpretation of the data assumes ous mutations (see Rice 1984; Patten and Haig 2009; Fry
that male-biased genes evolve by directional selection for male 2010). Concave fitness landscapes represent an interest-
expression divergence, and female-biased genes evolve by direc- ing and possibly widespread example of context-dependent
tional selection for female expression divergence. Our theoretical dominance, where deleterious alleles will be partially re-

results are valid whether this assumption holds true, yet interpre- cessive and beneficial alleles will be partially dominant.
tation of the empirical data, in light of the theory, is subject to Fry (2010) recently demonstrated that such dominance re-
revision pending future research. versals will accommodate an excess of balanced sexually
Given the empirical relationship between sex-biased gene antagonistic polymorphism on the autosomes relative to the
expression and sex-specific divergence described above, sex- X. Our study extends this result to show that, if sex-biased
specific selection theory makes five predictions about the genomic expression divergence involves the sequential invasion of
distribution of male- and female-biased genes sexually antagonistic alleles, followed by modifiers of sex-
limited expression divergence, sex-biased genes are likely
(1) When mutations underlying sex-specific expression diver- to accumulate more readily on autosomes (i.e., the invasion
gence have sex-limited and tissue-specific effects, genes conditions are more permissive, and the waiting time until
that are highly expressed in the heterogametic sex (males resolution is shorter).
in species with an X, females with a Z) will more read- (4) The relative rate of recombination in males versus females
ily evolve on autosomes. Genes that are highly expressed can have a major influence on the evolution and genomic
in the homogametic sex (XX females or ZZ males) may distribution of sex-biased genes. In species where recombi-
become preferentially sex linked when the effective muta- nation occurs in one sex only (i.e., the homogametic sex;
tion rate is relatively large per sex-linked locus (4NX uX > Haldane 1922; Huxley 1928; Lenormand and Dutheil 2005),
3NA uA or 4NZ uZ > 3NA uA )—a condition that is likely to the probability of invasion for sex-linked relative to autoso-
vary between different animal lineages (for recent reviews mal genetic combinations that cause sex-biased expression
of the theory and data, see Charlesworth 2009; Ellegren will decrease when they benefit the nonrecombining sex,
2009). This prediction has obvious parallels with those and increase when they benefit the recombining sex. When
of sex-linked versus autosome molecular evolution theory both sexes recombine, sex linkage will facilitate the evolu-
(see Kirkpatrick and Hall 2004; Vicoso and Charlesworth tion of both male- and female-biased gene expression. This
2009a). finding also provides a novel prediction for evolutionary
(2) Pleiotropy greatly accentuates the patterns predicted under theories of adaptive peak shifts, which had not previously
a model of sex- and tissue-limited mutation. Genes that are considered sex-linked inheritance (e.g., Crow and Kimura
highly expressed within the homogametic sex are expected 1965; Weinreich and Chao 2005; Kim 2007).
to accumulate at similar rates on the sex chromosomes and (5) The relative contribution of linkage and epistasis to ge-
autosomes, and may disproportionately accumulate on the nomic patterns of sexually dimorphic expression will ulti-
X or Z when 4NX uX > 3NA uA or 4NZ uZ > 3NA uA . Sex mately depend on the strength of evolutionary constraints
linkage severely constrains adaptive gene expression diver- influencing the evolution of interacting loci. As such, our
gence in the heterogametic sex, by reducing the range of results can be viewed along a continuum of evolutionary
parameters that are conducive to evolutionary divergence constraint. Under weak constraint, individual derived alle-
and by extending the mean waiting time until favored alleles can reach a high population frequency or become fixed in
les invade in the population. To our knowledge, this result the population. This initial invasion event can also generate
has not previously been reported in the population genetics selection for epistatic modifiers, and physical linkage be-
literature. tween epistatically interacting loci becomes less relevant
(3) Sexual antagonism over a gene’s expression level can rep- to the evolutionary dynamics. Under strong constraints,
resent a greater evolutionary constraint for sex-linked loci. antagonistically selected alleles persist as rare balanced

polymorphisms or as ephemeral deleterious mutations. behavior, mating receptivity and seminal fluid, and found that the
Coadaptation between these alleles and expression modi- number of X-linked genes was proportional to the size of the X
fiers is facilitated by, and in some cases may require, tight (as a function of euchromatin), suggesting a random chromoso-
physical linkage between the interacting loci. This suggests mal distribution of sexually selected genes. However, many genes
that regional genomic patterns of linkage and recombina- in the study were detected through their mutant phenotypes, and
tion will most strongly influence the evolution of sex-biased X-linked mutations are often easier to detect (e.g., Haldane 1935).
genes from strongly constrained precursors. Weakly con- Using a random set of visible mutations to establish a baseline
strained genes may show less sensitivity to physical linkage. expectation of X-linkage (Table A1 of Fitzpatrick 2004), the pro-
portion of sex-linked, sexually selected genes appears to be lower
SEX-SPECIFIC SELECTION AND THE OBSERVED than expected by chance (Table A1 set: 39% X-linked; sexually
CHROMOSOMAL DISTRIBUTION OF selected set: 27% X-linked). Because mutations at these sexually
SEX-BIASED GENES selected genes are generally pleiotropic, the pattern agrees with

Observed chromosomal distributions of sex-biased genes are com- the theoretical prediction regarding X-linkage and pleiotropy. In
patible with the theory outlined here. One of the most striking a human study using gene expression data from 14 different tis-
empirical patterns is the positive association between sex linkage sues, Lercher et al. (2003) found that expression breadth (as a
and preferential expression within the homogametic sex. Such metric of pleiotropy) was significantly lower for X-linked relative
linkage patterns have been reported in species of Caenorhab- to autosomal genes. An analysis of male-specific tissues (genes
ditis elegans, Drosophila, mouse, silkworm, and chicken (e.g., exclusively expressed in the prostate) enhances this bias toward
Ranz et al. 2003; Khil et al. 2004; Reinke et al. 2004; Kaiser and low pleiotropy on the X (again, as predicted by the theory).
Ellegren 2006; Arunkumar et al. 2009). This particular pattern is Two additional properties of X and Z chromosomes are
exclusively predicted by models of sex-specific selection, and can likely to influence linkage patterns of sex-biased genes. Sex chro-
potentially arise via multiple evolutionary pathways (e.g., through mosomes are inactivated during male meiosis in mammals and
invasion of sex-limited mutations and coevolution of epistatically Drosophila (Lifschytz and Lindsley 1972; Solari 1974; Hense
interacting, linked mutations). et al. 2007; Turner 2007), and during female meiosis in the chicken
Genes that are preferentially expressed within the heteroga- (Schoenmakers et al. 2009). There is now considerable evidence
metic sex exhibit inconsistent chromosomal distributions between from Drosophila that this process of meiotic sex chromosome
species, yet these lineage-specific patterns are both informa- inactivation (MSCI) constrains the evolution and retention of sex-
tive and compatible with models of sex-specific selection. In linked male-biased genes, particularly those expressed in testis
Drosophila, where males do not recombine, there is a deficit (Arbeitman et al. 2002; Bétran et al. 2002; Parisi et al. 2003; Wu
of X-linkage for genes with male-biased expression and male- and Xu 2003; Meisel et al. 2009; Vibranovski et al. 2009a,b). Con-
specific function (e.g., accessory gland proteins; Swanson et al. straints against sex-biased gene expression may also arise because
2001; Parisi et al. 2003; Mueller et al. 2005; Sturgill et al. 2007). the heterogametic sex carries a single copy of each X- or Z-linked
In birds and mammals, where recombination rates are relatively gene. To the extent that gene expression levels are constrained
similar between the sexes (Groenen et al. 2000; Stauss et al. 2003; by dosage compensation, highly expressed genes within the het-
Hansson et al. 2005; Lenormand and Dutheil 2005; Akesson et al. erogametic sex are expected to be disproportionately autosomal
2007; Backström et al. 2008; Hale et al. 2008; Stapley et al. 2008; (Rogers et al. 2003; Vicoso and Charlesworth 2006), a prediction
Jaari et al. 2009), genes that are preferentially expressed in the with some support from Drosophila (Vicoso and Charlesworth
heterogametic sex are enriched on the X or Z (Wang et al. 2001; 2009b; Bachtrog et al. 2010). Because these processes tend to
Lercher et al. 2003; Khil et al. 2004; Mořkovský et al. 2010). These skew male-biased genes toward autosomes, evolutionary con-
lineage-specific patterns suggest an important role for epistatic co- straint imposed by MSCI and dosage compensation will tend
evolution between linked cis-regulatory loci during the evolution to reinforce linkage patterns driven by sex-specific selection.
of sex-biased expression, and also agree nicely with widespread A distinct lack of nonrandomness between X and autosome
observations of cis-interactions mediating sex-specific regulation gene content has been reported for the mosquito Anopheles gam-
during development (Williams and Carroll 2009). biae (Hahn and Lanzaro 2005). This result may be shaped by
Whether interacting mutations are fixed sequentially or si- a suite of lineage-specific factors, including MSCI (reported to
multaneously, a negative relationship is expected between sex occur in Anopheles; McKee and Handel 1993), dosage compen-
linkage and the relative pleiotropy of male-biased genes. To date, sation (unconfirmed), as well as linkage and recombination. Re-
two studies have been published that are relevant to this prediction. combination occurs in male and female Anophelines—possibly
Fitzpatrick (2004) analyzed the chromosomal distribution of 63 at similar rates (Benedict et al. 2003)—which could effectively
Drosophila “sexually selected” genes associated with courtship depress the recombination rate per generation on the X and

generate a more favorable environment for male-biased genes scenario that is more in line with Rice’s (1984) classic analysis of
compared to Drosophila. On the other hand, sample size may sexually antagonistic alleles.
affect these conclusions, as relatively few genes included in the Although we focus on cis-regulatory substitutions, our re-
analysis are X-linked. Considering the entire dataset and using a sults can easily be generalized to incorporate a possible role of
twofold gene expression cut-off to define male- and female-biased trans-regulatory mutations during gene expression evolution. As
expression (M/F > 2 and M/F < 0.5, respectively; data obtained with cis-regulatory modifiers of sex- or tissue-limitation, the inva-
from the SEBIDA database: Gnad and Parsch 2006), A. gambiae sion of pleiotropic or sexually antagonistic alleles can potentially
shows a slight enrichment in female-biased genes (X = 8.3%; generate selection in favor of trans-acting modifier alleles (this is
n = 782 female-biased genes) and a deficit of male-biased genes similar to the scenario studied by Rice 1984). To the extent that
(X = 5.0%; n = 481 male-biased genes) relative to the genome- trans-acting modifiers are unconstrained by pleiotropy, sequential
wide expectation (X = 7.1% for the entire dataset; n = 4281 total coevolution between cis- and trans-alleles may be likely. This also
genes). However, given the low proportion of X-linkage, neither opens up the possibility of epistatic coevolution between loci on

difference is statistically significant. different chromosomes (including coevolution between the X, Y,
and autosomes), which may be important during the evolution of
THE SHAPE OF THE EXPRESSION-FITNESS reproductive isolation between species (see Ortiz-Barrientos et al.
LANDSCAPE 2007). However, simultaneous invasion of positively interacting
Predictions of the theory depend on the shape of the relationship cis- and trans-mutations will generally be impossible if such muta-
between gene expression and fitness, which we have assumed fol- tions are not closely linked. Thus, cis-/cis-epistasis should permit
lows a concave function, at least within the region of trait space a broader range of evolutionary transitions relative to cis-/trans-
where the evolution of sexually dimorphic expression occurs. If or trans-/trans-interactions. Because trans-regulatory mutations
most gene expression, evolution occurs within the vicinity of a are also expected to be more pleiotropic, they should provide little
local fitness optimum, then the assumption of fitness concavity opportunity for adaptation compared to cis-regulatory mutations
does not appear to be controversial, and indeed, several lines of (Stern 2000; Prud’homme et al. 2007; Wray 2007). Nevertheless,
empirical and theoretical evidence support it (e.g., Wright 1934; current genomic analyses of sex-biased gene expression cannot
Charlesworth 1979; Kacser and Burns 1981; Dekel and Alon rule out a role of trans-substitutions during the evolutionary ori-
2005; Lunzer et al. 2005; Phadnis and Fry 2005; Kalisky et al. gin of sex-biased expression, and future empirical research will
2007; Bedford and Hartl 2009; Zhang et al. 2009; see above). be required to further address this issue.
This assumption is further justified by analogy to mathematical
theories of DNA adaptation, which suggest that beneficial substi-
tutions will follow a diminishing returns function (Gillespie 1984; Conclusion
see Orr 2005 for a clear discussion of this body of work). The spe- Sex-specific selection, particularly selection for gene expression
cific degree of fitness concavity, which we incorporate into our divergence between males and females, can play a prominent
model with parameter k (where the fitness function is concave role in shaping the genomic distributions of sex-biased genes and
whenever k > 1, and the degree of concavity increases with k), genes underlying sexual dimorphism. We show that, although
cannot currently be inferred empirically. We simply note that the the underlying patterns of selection can be complex, the genomic
actual value of k has little impact on qualitative predictions of predictions of sex-specific selection hypotheses are generally con-
the models as long as k > 1 (see Figures S3–S6 for analyses sistent with empirical distributions of sex-biased genes, although
with different values of k). Thus, to the extent that selection for the current data apply to a relatively small (but growing) set of
sex-specific divergence occurs on a concave expression-fitness well-characterized species. Future genome-wide analyses of sex-
function, our qualitative predictions will be robust to the actual specific genetic architecture, including the distributions of genes
degree of concavity. with sex-specific expression patterns in nonmodel organisms, will
The concavity assumption may be violated for genes that, permit a better evaluation of these theoretical predictions.
prior to adaptation, are severely distant from a local fitness opti- One notable pattern emerging from this theory is that chro-
mum. Although concavity near the optimum is expected, linear mosomal distributions of sex-biased genes are likely to have both
(k = 1) or convex (0 < k < 1), fitness landscapes could potentially gene- and species-specific attributes. These predictions may help
arise within other regions of trait space. Given our analysis above, to inform future genomic analyses, and may point toward the
convex fitness surfaces will cause a shift toward recessive fitness specific evolutionary processes underlying a diversity of linkage
effects of beneficial mutations (in terms of our model, hb < 0.5 patterns between gene categories and species.
when 0 < k < 1). Under this condition, the probability of invasion Finally, we focus on selection acting on gene expression vari-
for a male-beneficial mutation will be enhanced by X-linkage, a ation in an attempt to explain the evolution of a gene expression

phenotype (i.e., sex-biased gene expression) that is ubiquitous Benedict, M. Q., L. M. McNitt, and F. H. Collins. 2003. Genetic traits of the
at the molecular genomic scale. Intersexual divergence in gene mosquito Anopheles gambiae: red stripe, frizzled, and homochromy1.
J Hered. 94:227–235.
expression may play a general role during the evolution of sex-
Bétran, E., K. Thornton, and M. Long. 2002. Retroposed genes out of the X
ually dimorphic phenotypes, including sex-specific morphology, in Drosophila. Genome Res. 12:1854–1859.
behavior, and life history. On the other hand, the genetic basis Blackburn, G. S., A. Y. K. Albert, and S. P. Otto. 2010. The evolution of sex
of sexual dimorphism may critically depend on the types of ge- ratio adjustment in the presence of sexually antagonistic selection. Am.
netic substitutions that are differentially favored between males Nat. 176:264–275.
Bonduriansky, R., and S. F. Chenoweth. 2009. Intralocus sexual conflict.
and females. In principle, disruptive selection might favor protein Trends Ecol. Evol. 24:280–288.
sequence divergence between the sexes, yet it is difficult to see Calsbeek, R., and C. Bonneaud. 2008. Postcopulatory fertilization bias as a
how conflicts over coding sequence would necessarily become form of cryptic sexual selection. Evolution 62:1137–1148.
resolved by the evolution of dimorphic gene expression alone. Candolin, U., and J. Heuschele. 2008. Is sexual selection beneficial during
adaptation to environmental change? Trends Ecol. Evol. 23:446–452.
Additional theory, particularly theory involving the evolution of

Carroll, S. B. 2008. Evo-devo and an expanding evolutionary synthesis: a
alternative splicing and gene duplication, will be required to ad- genetic theory of morphological evolution. Cell 134:25–36.
dress how sexual dimorphism might evolve from an initial condi- Charlesworth, B. 1979. Evidence against Fisher’s theory of dominance. Nature
tion of coding sequence conflict between the sexes. Experimental 278:848–849.
———. 2001. The effect of life-history and mode of inheritance on neutral
research will also be required to better understand the degree to
genetic variability. Genet. Res. 77:153–166.
which coding and noncoding mutations contribute to sex-specific ———. 2009. Effective population size and patterns of molecular evolution
fitness variation (for a promising step toward this difficult goal, and variation. Nat. Rev. Genet. 10:195–205.
see Innocenti and Morrow 2010). Charlesworth, B., J. A. Coyne, and N. H. Barton. 1987. The relative rates
of evolution of sex chromosomes and autosomes. Am. Nat. 130:113–
146.
ACKNOWLEDGMENTS Chintapalli, V. R., J. Wang, and J. A. T. Dow. 2007. Using FlyAtlas to identify
This work benefited greatly from discussions with members of the Clark better Drosophila models of human disease. Nat. Genet. 39:715–720.
Lab, especially R. Meisel, and comments from three anonymous review- Clark, A. G. 1991. Mutation-selection balance and metabolic control theory.
ers. Funding was provided by an NIH grant (GM64590) to AGC and A. Genetics 129:909–923.
Bernardo Carvalho. Connallon, T. 2010. Genic capture, sex-linkage, and the heritability of fitness.
Am Nat. 175:564–576.
LITERATURE CITED Connallon, T., and L. L. Knowles. 2005. Intergenomic conflict revealed by
Agrawal, A. F. 2001. Sexual selection and the maintenance of sexual repro- patterns of sex-biased gene expression. Trends Genet. 21:495–499.
duction. Nature 411:692–695. Connallon, T., R. M. Cox, and R. Calsbeek. 2010. Fitness consequences of
Akesson, M., B. Hansson, D. Hasselquist, and S. Bensch. 2007. Linkage sex-specific selection. Evolution 64:1671–1682.
mapping of AFLP markers in a wild population of great reed warblers: Cox, R. M., and R. Calsbeek. 2009. Sexually antagonistic selection, sexual
importance of heterozygosity and number of genotyped individuals. dimorphism, and the resolution of intralocus sexual conflict. Am. Nat.
Mol. Ecol. 16:2189–2202. 173:176–187.
Albert, A. Y. K., and S. P. Otto. 2005. Sexual selection can resolve sex-linked ———. 2010. Cryptic sex-ratio bias provides indirect genetic benefits despite
sexual antagonism. Science 310:119–121. sexual conflict. Science 328:92–94.
Andersson, M. 1994. Sexual selection. Princeton Univ. Press, NJ. Coyne, J. A., E. H. Kay, and S. Pruett-Jones. 2008. The genetic basis of sexual
Arbeitman, M. N., E. E. M. Furlong, F. Imam, E. Johnson, B. H. Null, dimorphism in birds. Evolution 62:214–219.
B. S. Baker, M. A. Krasnow, M. P. Scott, R. W. Davis, and K. P. White. Crow, J. F., and M. Kimura. 1965. Evolution in sexual and asexual populations.
2002. Gene expression during the life cycle of Drosophila melanogaster. Am. Nat. 99:439–450.
Science 297:2270–2275. ———-. 1970. An introduction to population genetics theory. Harper & Row,
Arqvist, G., and L. Rowe. 2005. Sexual conflict. Princeton Univ. Press, New York, NY.
Princeton, NJ. Curtsinger, J. W., P. M. Service, and T. Prout. 1994. Antagonistic pleiotropy,
Arunkumar, K. P., K. Mita, and J. Nagaraju. 2009. The silkworm Z chromo- reversal of dominance, and genetic polymorphism. Am. Nat. 144:210–
some is enriched in testis-specific genes. Genetics 182:493–501. 228.
Bachtrog, D., N. R. T. Toda, and S. Lockton. 2010. Dosage compensation Darwin, C. 1871. The descent of man, and selection in relation to sex. John
and demasculinization of X chromosomes in Drosophila. Curr. Biol. Murray, London.
20:1476–1481. Dean, A. M. 1989. Selection and neutrality in lactose operons of Escherichia
Backström, N., N. Karaiskou, E. H. Leder, L. Gustafsson, C. R. Primmer, A. coli. Genetics 123:441–454.
Qvarnström, and H. Ellegren. 2008. A gene-based genetic linkage map Dean, A. M., D. E. Dykhuizen, and D. L. Hartl. 1986. Fitness as a function of
of the collared flycatcher (Ficedula albicollis) reveals extensive synteny beta-galactosidase activity in Escherichia coli. Genet. Res. 48:1–8.
and gene order conservation during 100 million years of avian evolution. Dekel, E., and U. Alon. 2005. Optimality and evolutionary tuning of the
Genetics 179:1479–1495. expression level of a protein. Nature 436:588–592.
Bedford, T., and D. L. Hartl. 2009. Optimization of gene expression by natural Dykhuizen, D. E., A. M. Dean, and D. L. Hartl. 1987. Metabolic flux and
selection. Proc. Natl. Acad. Sci. USA 106:1133–1138. fitness. Genetics 115:25–31.
Bell, G. 1982. The masterpiece of nature: the evolution and genetics of sexu- Ellegren, H. 2007. Characteristics, causes and evolutionary consequences of
ality. Univ. of California Press, Berkeley, CA. male-biased mutation. Proc. Biol. Sci. 274:1–10.

———. 2009. The different levels of genetic diversity in sex chromosomes Hurst, L. D., and J. P. Randerson. 2000. Dosage, deletions and dominance: sim-
and autosomes. Trends Genet. 25:278–284. ple models of the evolution of gene expression. J. Theor. Biol. 205:641–
Ellegren, H., and J. Parsch. 2007. The evolution of sex-biased genes and 647.
sex-biased gene expression. Nat. Rev. Genet. 8:689–698. Huxley, J. 1928. Sexual difference of linkage in Gammarus chevreuxi.
Fedorka, K. M., and T. A. Mousseau. 2004. Female mating bias results in J. Genet. 20:145–156.
conflicting sex-specific offspring fitness. Nature 429:65–67. Innocenti, P., and E. H. Morrow. 2010. The sexually antagonistic genes of
Fisher, R. A. 1958. The genetical theory of natural selection, 2nd ed. Dover Drosophila melanogaster. PLoS Biol. 8:e1000335.
Publications, Inc., New York. Jaari, S., L. Meng-Hua, and J. Merilä. 2009. A first-generation microsatellite-
Fitzpatrick, M. J. 2004. Pleiotropy and the genomic location of sexually based genetic linkage map of the Siberian jay (Perisoreus infaustus):
selected genes. Am. Nat. 163:800–808. insights into avian genome evolution. BMC Genomics 10:1.
Fry, J. D. 2010. The genomic location of sexual antagonistic variation: some Kacser, H., and J. A. Burns. 1981. The molecular basis of dominance. Genetics
cautionary comments. Evolution 64:1510–1516. 97:639–666.
Garcı́a-Dorado, A., A. Caballero, and J. F. Crow. 2003. On the persistence Kacser, H., and R. Beeby. 1984. Evolution of catalytic proteins or on the
and pervasiveness of a new mutation. Evolution 57:2644–2646. origin of enzyme species by means of natural selection. J. Mol. Evol.

Gillespie, J. H. 1984. Molecular evolution over the mutational landscape. 20:38–51.
Evolution 38:1116–1129. Kaiser, V. B., and H. Ellegren. 2006. Nonrandom distribution of genes with
———. 1991. The causes of molecular evolution. Oxford Univ. Press, Oxford, sex-biased expression in the chicken genome. Evolution 60:1945–1951.
UK. Kalisky, T., E. Dekel, and U. Alon. 2007. Cost-benefit theory and optimal
———. 2000. Genetic drift in an infinite population: the pseudohitchhiking design of gene regulatory functions. Phys. Biol. 4:229–245.
model. Genetics 155:909–919. Khil, P. P., N. A. Smirnova, P. J. Romanienko, and R. D. Camerini-Otero.
Gnad, F., and J. Parsch. 2006. Sebida: a database for the functional and evo- 2004. The mouse X chromosome is enriched for sex-biased genes not
lutionary analysis of genes with sex-biased expression. Bioinformatics subject to selection by meiotic sex chromosome inactivation. Nat. Genet.
22:2577–2579. 36:642–646.
Gout, J.-F., D. Kahn, L. Duret, and Paramecium Post-Genomics Consortium. Kidwell, J. F., M. T. Clegg, F. M. Stewart, and T. Prout. 1977. Regions of
2010. The relationship among gene expression, the evolution of dosage, stable equilibria for models of differential selection in two sexes under
and the rate of protein evolution. PLoS Genet. 6:e1000944. random mating. Genetics 85:171–183.
Graze, R. M., L. M. McIntyre, B. J. Main, M. L. Wayne, and S. V. Nuzhdin. Kim, Y. 2007. Rate of adaptive peak shifts with partial genetic robustness.
2009. Regulatory divergence in Drosophila melanogaster and D. sim- Evolution 61:1847–1856.
ulans, a genomewide analysis of allele-specific expression. Genetics Kirkpatrick, M., and D. W. Hall. 2004. Male-biased mutation, sex linkage,
183:547–561. and the rate of adaptive evolution. Evolution 58:437–440.
Groenen, M. A., H. H. Cheng, N. Bumstead, B. F. Benkel, W. E. Briles, Koeslag, P. D., and J. H. Koeslag. 1994. Koinophilia stabilizes bi-gender sexual
T. Burke, D. W. Burt, L. B. Crittenden, J. Dodgson, J. Hillel, et al. reproduction against asex in an unchanging environment. J. Theor. Biol.
2000. A consensus linkage map of the chicken genome. Genome Res. 166:251–260.
10:137–147. Kondrashov, A. S. 1988. Deleterious mutations as an evolutionary factor. III.
Gurbich, T. A., and D. Bachtrog. 2008. Gene content evolution on the X Mating preferences and some general remarks. J. Theor. Biol. 131:487–
chromosome. Curr. Opin. Genet. Dev. 18:493–498. 496.
Hadany, L., and T. Beker. 2007. Sexual selection and the evolution of obliga- Lande, R. 1980. Sexual dimorphism, sexual selection, and adaptation in poly-
tory sex. BMC Evol. Biol. 7:e245. genic characters. Evolution 34:292–305.
Hahn, M. W., and G. C. Lanzaro. 2005. Female-biased expression in the Lenormand, T., and J. Dutheil. 2005. Recombination differences between
malaria mosquito Anopheles gambiae. Curr, Biol. 15:R192–R193. sexes: a role for haploid selection. PLoS Biol. 3:396–403.
Haldane, J. B. S. 1922. Sex ration and unisexual sterility in hybrid animals. Lercher, J. M., A. O. Urrutia, and L. D. Hurst. 2003. Evidence that the human
J. Genet. 12:101–109. X chromosome is enriched for male-specific but not female-specific
———. 1927. A mathematical theory of natural and artificial selection. genes. Mol. Biol. Evol. 20:1113–1116.
Part V: Selection and mutation. Proc. Camb. Phil. Soc. 23:838–844. Lifschytz, E., and D. L. Lindsley. 1972. The role of X chromosome inac-
———. 1935. The rate of spontaneous mutation of a human gene. J. Genet. tivation during spermatogenesis. Proc. Natl. Acad. Sci. USA 69:182–
31:317–326. 186.
———. 1962. Conditions for stable polymorphism at an autosomal locus. Lion, S., F. Gabriel, B. Bost, J. Fiévet, C. Dillmann, and D. de Vienne. 2004.
Nature 193:1108. An extension to the metabolic control theory taking into account cor-
Hale, M. C., H. Jensen, T. R. Birkhead, T. Burke, and J. Slate. 2008. A relations between enzyme concentrations. Eur. J. Biochem. 271:4375–
comparison of synteny and gene order on the homologue of chicken 4391.
chromosome 7 between two passerine species and between passerines Loehlin, D. W., L. S. Enders, and J. H. Werren. 2010a. Evolution of sex-
and chicken. Cyt. Gen. Res. 121:120–129. specific wing shape at the widerwing locus in four species of Nasonia.
Hansson, B., M. Akesson, J. Slate, and J. M. Pemberton. 2005. Linkage Heredity 104:260–269.
mapping reveals sex-dimorphic map distances in a passerine bird. Proc. Loehlin, D. W., D. C. S. G. Oliveira, R. Edwards, J. D. Giebel, M. E. Clark,
R. Soc. Lond. B 272:2289–2298. M. V. Cattani, L. van de Zande, E. C. Verhulst, L. W. Beukeboom,
Hartl, D. L., D. E. Dykhuizen, and A. M. Dean. 1985. Limits on adaptation: M. Muñoz-Torres, et al. 2010b. Non-coding changes cause sex-specific
the evolution of selective neutrality. Genetics 111:655–674. wing size differences between closely related species of Nasonia. PLoS
Hedrick, P. W. 2007. Sex: differences in mutation, recombination, selection, Genet. 6:e1000821.
gene flow, and genetic drift. Evolution 61:2750–2771. Lorch, P. D., S. Proulx, L. Rowe, and T. Day. 2003. Condition-dependent
Hense, W., J. F. Baines, and J. Parsch. 2007. X chromosome inactivation sexual selection can accelerate adaptation. Evol. Ecol. Res. 5:867–
during Drosophila spermatogenesis. PLoS Biol. 5:e273. 881.

Lunzer, M., S. P. Milter, R. Felsheim, and A. M. Dean. 2005. The biochemical Patten, M. M., and D. Haig. 2009. Maintenance or loss of genetic variation
architecture of an ancient adaptive landscape. Science 310:499–501. under sexual and parental antagonism at a sex-linked locus. Evolution
Mackay, T. F. C. 2001. The genetic architecture of quantitative traits. Annu. 63:2888–2895.
Rev. Genet. 35:303–339. Phadnis, N., and J. D. Fry. 2005. Widespread correlations between domi-
Mank, J. E. 2009. Sex chromosomes and the evolution of sexual dimorphism: nance and homozygous effects of mutations: implications for theories
lessons from the genome. Am. Nat. 173:141–150. of dominance. Genetics 171:385–392.
Mank, J. E., and H. Ellegren. 2009. Sex linkage of sexually antagonistic Pischedda, A., and A. K. Chippindale. 2006. Intralocus sexual conflict dimin-
genes is predicted by female, but not male, effects in birds. Evolution ishes the benefits of sexual selection. PLoS Biol. 4:2099–2103.
63:1464–1472. Poissant, J., A. J. Wilson, and D. W. Coltman. 2010. Sex-specific genetic
Mank, J. E., L. Hultin-Rosenberg, M. Zwahlen, and H. Ellegren. 2008. variance and the evolution of sexual dimorphism: a systematic review
Pleiotropic constraint hamper the resolution of sexual antagonism in of cross-sex genetic correlations. Evolution 64:97–107.
vertebrate gene expression. Am. Nat. 11:35–43. Prout, T. 1999. How well does opposing selection maintain variation?
Manning, J. T. 1984. Males and the advantage of sex. J. Theor. Biol. 108:215– Pp. 157–182 in R. C. Lewontin, R. S. Singh, C. B. Krimbas, and
220. K. V. Krimpas, eds. Evolutionary genetics: from molecules to mor-

McKee, B. D., and M. A. Handel. 1993. Sex chromosomes, recombination, phology. Cambridge Univ. Press, Boston.
and chromatin conformation. Chromosoma 102:71–80. Prud’homme, B., N. Gompel, and S. B. Carroll. 2007. Emerging princi-
Meisel, R. P., M. V. Han, and M. W. Hahn. 2009. A complex suite of forces ples of regulatory evolution. Proc. Natl. Acad. Sci. USA 104:8605–
drives gene traffic from Drosophila X chromosomes. Genome Biol. 8612.
Evol. 1:176–188. Ranz, J. M., C. I. Castillo-Davis, C. D. Meiklejohn, and D. L. Hartl. 2003.
Middleton, R. J., and H. Kacser. 1983. Enzyme variation, metabolic flux and Sex-dependent gene expression and evolution of the Drosophila tran-
fitness: alcohol dehydrogenase in Drosophila melanogaster. Genetics scriptome. Science 300:1742–1745.
105:633–650. Reinke, V., H. E. Smith, J. Nance, J. Wang, C. Van Doren, R. Begley, S.
Mořkovský, L., R. Storchova, J. Plachy, R. Ivanek, P. Divina, and J. Hejnar. J. M. Jones, E. B. Davis, S. Scherer, S. Ward, et al. 2000. A global
2010. The chicken Z chromosome is enriched for genes with preferential profile of germ line gene expression in C. elegans. Mol. Cell. 6:605–
expression in ovarian somatic cells. J. Mol. Evol. 70:129–136. 616.
Morrow, E. H., A. D. Stewart, and W. R. Rice. 2008. Assessing the extent Reinke, V., I. S. Gil, S. Ward, and K. Kazmer. 2004. Genome-wide germline-
of genome-wide intralocus sexual conflict via experimentally enforced enriched and sex-biased expression profiles in Caenorhabditis elegans.
gender-limited selection. J. Evol. Biol. 21:1046–1054. Development 131:311–323.
Mueller, J. L., K. Ravi Ram, L. A. McGraw, M. C. Bloch Qazi, E. D. Siggia, Rhen, T. 2000. Sex-limited mutations and the evolution of sexual dimorphism.
A. G. Clark, C. F. Aquadro, and M. F. Wolfner. 2005. Cross-species Evolution 54:37–43.
comparison of Drosophila male accessory gland protein genes. Genetics Rice, W. R. 1984. Sex chromosomes and the evolution of sexual dimorphism.
171:131–143. Evolution 38:735–742.
Mueller, J. L., S. K. Mahadevaiah, P. J. Park, P. E. Warburton, D. C. Page, and ———. 2002. Experimental tests of the adaptive significance of sexual re-
J. M. A. Turner. 2008. The mouse X chromosome is enriched for multi- combination. Nat. Rev. Genet. 3:241–251.
copy testis genes showing postmeiotic expression. Nat. Genet. 40:794– Rice, W. R., and A. K. Chippindale. 2001. Intersexual ontogenetic conflict.
799. J. Evol. Biol. 14:685–693.
Ohta, T. 1992. The nearly neutral theory of molecular evolution. Annu. Rev. Rogers, D. W., M. Carr, and A. Pomiankowski. 2003. Male genes: X-pelled
Ecol. Syst. 23:263–286. or X-cluded? Bioessays 25:739–741.
Oliver, B., and M. Parisi. 2004. Battle of the Xs. Bioessays 26:543–548. Schoenmakers, S., E. Wassenaar, J. W. Hoogerbrugge, J. S. E. Laven, J. A.
Orr, H. A. 1991. A test of Fisher’s theory of dominance. Proc. Natl. Acad. Grootegoed, and W. M. Baarends. 2009. Female meiotic sex chromo-
Sci. USA 88:11413–11415. some inactivation in chicken. PLoS Genet. 5:e10000466.
———. 2005. The genetic theory of adaptation: a brief history. Nat. Rev. Scotti, I., and L. F. Delph. 2006. Selective trade-offs and sex-chromosome
Genet. 6:119–127. evolution in Silene latifolia. Evolution 60:1793–1800.
Ortı́z-Barrientos, D., B. A. Counterman, and M. A. F. Noor. 2007. Genet Siller, S. 2001. Sexual selection and the maintenance of sex. Nature 411:689–
expression divergence and the origin of hybrid dysfunctions. Genetica 692.
129:71–81. Solari, A. J. 1974. The behavior of the XY pair in mammals. Int. Rev. Cytol.
Otto, S. P., and D. Bourguet. 1999. Balanced polymorphism and the evolution 38:273–317.
of dominance. Am. Nat. 153:561–574. Stapley, J., T. R. Birkhead, T. Burke, and J. Late. 2008. A linkage map of
Otto, S. P., and T. Day. 2007. A biologist’s guide to mathematical modeling the zebra finch Taeniopygia guttata provides new insights into avian
in ecology and evolution. Princeton Univ. Press, NJ. genome evolution. Genetics 179:651–667.
Otto, S. P., and P. Yong. 2002. The evolution of gene duplicates. Homol. Stauss, M., J. Tomiuk, G. Segelbacher, S. Driesel, J. Fietz, L. Bachmann,
Effects 46:451–483. and J. Kompf. 2003. Sex-specific recombination rates in Parus major
Owen, A. R. G. 1953. A genetical system admitting of two distinct stable and P. caeruleus, an exception to Huxley’s rule. Hereditas 139:199–
equilibria under natural selection. Heredity 7:97–102. 205.
Pamilo, P. 1979. Genic variation at sex-linked loci: quantification of regular Stephan, W. 1996. The rate of compensatory evolution. Genetics 144:419–
selection models. Hereditas 91:129–133. 426.
Papp, B., C. Pal, and L. D. Hurst. 2003. Dosage sensitivity and the evolution Stern, D. L. 2000. Evolutionary developmental biology and the problem of
of gene families in yeast. Nature 424:194–197. variation. Evolution 54:1079–1091.
Parisi, M., R. Nuttall, D. Naiman, G. Bouffard, J. Malley, J. Andrews, S. Stewart, A. D., A. Pischedda, and W. R. Rice. 2010. Resolving intralocus
Eastman, and B. Oliver. 2003. Paucity of genes on the Drosophila X sexual conflict: genetic mechanisms and timeframe. J. Hered. 101:S94–
chromosome showing male-biased expression. Science 299:697–700. S99.

Storchova, R., and P. Divina. 2006. Nonrandom representation of sex- Whitlock, M. C., and A. F. Agrawal. 2009. Purging the genome with sexual
biased genes on chicken Z chromosome. J. Mol. Evol. 63:676– selection: reducing mutation load through selection in males. Evolution
681. 63:569–582.
Sturgill, D., Y. Zhang, M. Parisi, and B. Oliver. 2007. Demasculiniza- Williams, T. M., and S. B. Carroll. 2009. Genetic and molecular insights
tion of X chromosomes in the Drosophila genus. Nature 450:238– into the development and evolution of sexual dimorphism. Nat. Rev.
241. Genet.10;797–804.
Swanson, W. J., A. G. Clark, H. M. Waldrip-Dail, M. F. Wolfner, and C. F. Wittkopp, P. J., B. K. Haerum, and A. G. Clark. 2004. Evolutionary changes
Aquadro. 2001. Evolutionary EST analysis identifies rapidly evolving in cis and trans gene regulation. Nature 430:85–88.
male reproductive proteins in Drosophila. Proc. Natl. Acad Sci. USA ———. 2008. Regulatory changes underlying expression differences within
93:7375–7379. and between Drosophila species. Nat. Genet. 40:346–350.
Takahasi, K. R., and F. Tajima. 2005. Evolution of coadaptation in a two-locus Wray, G. A. 2007. The evolutionary significance of cis-regulatory mutations.
epistatic system. Evolution 59:2324–2332. Nat. Rev. Genet. 8:206–216.
Trivers, R. L. 1972. Parental investment and sexual selection. Pp. 136–179 in Wright, S. 1934. Physiological and evolutionary theories of dominance. Am.
B. Campbell, ed. Sexual selection and the descent of man, 1871–1971. Nat. 68:24–53.

Aldine, Chicago. Wu, C. I., and E. Y. Xu. 2003. Sexual antagonism and X inactivation—the
Turner, J. M. A. 2007. Meiotic sex chromosome inactivation. Development SAXI hypothesis. Trends Genet. 19:243–247.
134:1823–1831. Wyman, M. J., A. F. Agrawal, and L. Rowe. 2010. Condition-dependence
van Doorn, G. S. 2009. Intralocus sexual conflict. Annals NY Acad. Sci. of the sexually dimorphic transcriptome in Drosophila melanogaster.
1168:52–71. Evolution 64:1836–1848.
Vibranovski, M. D., H. F. Lopes, T. L. Karr, and M. Y. Long. 2009a. Stage- Zhang, Z., W. Qian, and J. Zhang. 2009. Positive selection for elevated gene
specific expression profiling of Drosophila spermatogenesis suggests expression noise in yeast. Mol. Syst. Biol. 5:299.
that meiotic sex chromosome inactivation drives genomic relocalization
of testis-expressed genes. PLoS Genet. 5:e10000731. Associate Editor: K. Dyer
Vibranovski, M. D., Y. Zhang, and M. Y. Long. 2009b. General movement
off the X chromosome in the Drosophila genus. Genome Res. 19:897–
903. Appendix 1: Development of
Vicoso, B., and B. Charlesworth. 2006. Evolution on the X chromosome:
unusual patterns and processes. Nat. Rev. Genet. 7:645–653.
The Recursions
———. 2009a. Effective population size and the faster-X effect: an extended
Haplotypes and sex-specific frequencies
model. Evolution 63:2413–2426.
———. 2009b. The deficit of male-biased genes on the D. melanogaster X
Haplotype Frequency in Sperm Frequency in Eggs
chromosome is expression-dependent: a consequence of dosage com-
pensation? J. Mol. Evol. 68:576–583. A1 B1 y1 x1
Wagner, A. 2005. Energy constraints on the evolution of gene expression.
A1 B2 y2 x2
Mol. Biol. Evol. 22:1365–1374.
A2 B1 y3 x3
———. 2007. Energy costs constrain the evolution of gene expression.
A2 B2 y4 x4
J. Exper. Zool. 308B:322–324.
Wang, P. J., J. R. McCarrey, F. Yang, and D. C. Page. 2001. An abundance
of X-linked genes expressed in spermatogonia. Nat. Genet. 27:422– Following the convention of much of the sex-specific selec-
426.
tion literature of tallying allele (or in our case haplotype) frequen-
Weinreich, D. M., and L. Chao. 2005. Rapid evolutionary escape by large pop-
ulations from local fitness peaks is likely in nature. Evolution 59:1175– cies within the eggs and sperm (e.g., Owen 1953; Kidwell et al.
1182. 1977; Patten and Haig 2009), we developed the following recur-
Whitlock, M. C. 2000. Fixation of new alleles and the extinction of small sion equations for the autosomes. Expected frequencies of each
populations: drift load, beneficial alleles, and sexual selection. Evolution
haplotype within male and female gametes in the next generation
54:1855–1861.
are
2x1 y1 m 11 + (x1 y2 + x2 y1 )m 12 + (x1 y3 + x3 y1 )m 21 + (1 − rm )m 22C (x1 y4 + x4 y1 ) + rm m 22R (x2 y3 + x3 y2 )

y1 =
2w̄m
2x2 y2 m 13 + (x1 y2 + x2 y1 )m 12 + (x2 y4 + x4 y2 )m 23 + (1 − rm )m 22R (x3 y2 + x2 y3 ) + rm m 22C (x4 y1 + x1 y4 )
y2 =
2w̄m
2x3 y3 m 31 + (x1 y3 + x3 y1 )m 21 + (x3 y4 + x4 y3 )m 32 + (1 − rm )m 22R (x2 y3 + x3 y2 ) + rm m 22C (x4 y1 + x1 y4 )
y3 =
2w̄m
2x4 y4 m 33 + (x2 y4 + x4 y2 )m 23 + (x4 y3 + x3 y4 )m 32 + (1 − rm )m 22C (x4 y1 + x1 y4 ) + rm m 22R (x2 y3 + x3 y2 )
y4 =
2w̄m
2x1 y1 f 11 + (x1 y2 + x2 y1 ) f 12 + (x1 y3 + x3 y1 ) f 21 + (1 − r f ) f 22C (x1 y4 + x4 y1 ) + r f f 22R (x2 y3 + x3 y2 )
x1 =
2w̄ f

2x2 y2 f 13 + (x1 y2 + x2 y1 ) f 12 + (x2 y4 + x4 y2 ) f 23 + (1 − r f ) f 22R (x2 y3 + x3 y2 ) + r f f 22C (x4 y1 + x1 y4 )

x2 =
2w̄ f

x3 =
2w̄ f

x4 = .
2w̄ f
Within a given generation, mean fitness in males and females,
respectively, at the autosomal loci will be w̄ f = x1 y1 f 11 + (x1 y2 + x2 y1 ) f 12 + x2 y2 f 13
+ (x1 y3 + x3 y1 ) f 21 + (x2 y3 + x3 y2 ) f 22R
w̄m = x1 y1 m 11 + (x1 y2 + x2 y1 )m 12 + x2 y2 m 13 + (x1 y4 + x4 y1 ) f 22C + (x2 y4 + x4 y2 ) f 23

+ (x1 y3 + x3 y1 )m 21 + (x2 y3 + x3 y2 )m 22R + x3 y3 f 31 + (x3 y4 + x4 y3 ) f 32 + x4 y4 f 33 .
+ (x1 y4 + x4 y1 )m 22C + (x2 y4 + x4 y2 )m 23
+ x3 y3 m 31 + (x3 y4 + x4 y3 )m 32 + x4 y4 m 33
Appendix 2: Stability with A1
w̄ f = x1 y1 f 11 + (x1 y2 + x2 y1 ) f 12 + x2 y2 f 13 and B1 Fixed
+ (x1 y3 + x3 y1 ) f 21 + (x2 y3 + x3 y2 ) f 22R To evaluate stability of the equilibrium x1 = y1 = 1, we calculated
+ (x1 y4 + x4 y1 ) f 22C + (x2 y4 + x4 y2 ) f 23 the following Jacobian matrices for the system of X-linked and
+ x3 y3 f 31 + (x3 y4 + x4 y3 ) f 32 + x4 y4 f 33 . autosomal recursions. These matrices, evaluated at the aforemen-
tioned state (x1 = y1 = 1), and rearranged to block-triangular
Recursions for the X chromosome (for Z chromosome recur-
form, are
sions, reverse males and females) are given by ⎛ ∂x ∂x ∂x ∂x ∂x ∂x
⎞
0 0 ∂ x12 ∂ y21 ∂ x13 ∂ y31 ∂ x14 ∂ y41
x1 m 1 ⎜ ⎟
y1 = ⎜ 0 0 ∂ y1 ∂ y1 ∂ y1 ∂ y1 ∂ y1 ∂ y1 ⎟
w̄m ⎜ ∂ x2 ∂ y2 ∂ x3 ∂ y3 ∂ x4 ∂ y4 ⎟
⎜ ⎟
⎜ 0 0 ∂ x2 ∂ x2 0 0
∂ x2 ∂ x2 ⎟
⎜ ∂ x2 ∂ y2 ∂ x4 ∂ y4 ⎟
x2 m 2 ⎜ ⎟
y2 = ⎜ 0 0 ∂ y2 ∂ y2 0 0
∂ y2 ∂ y2 ⎟
w̄m ⎜ ∂ x2 ∂ y2 ∂ x4 ∂ y4 ⎟
J A |x1 =y1 =1 = ⎜ ∂ x3 ∂ x3 ∂ x3 ∂ x3 ⎟
⎜0 0 0 0 ⎟
⎜ ∂ x3 ∂ y3 ∂ x4 ∂ y4 ⎟
⎜ ∂ y3 ∂ y3 ∂ y3 ∂ y3 ⎟
x3 m 3 ⎜0 0 0 0 ⎟
y3 = ⎜ ∂ x3 ∂ y3 ∂ x4 ∂ y4 ⎟
w̄m ⎜ ∂ x4 ∂ x4 ⎟
⎜0 0 0 0 0 0 ∂ x4 ∂ y4 ⎟
⎝ ⎠
∂ y4 ∂ y4
x4 m 4 0 0 0 0 0 0 ∂ x4 ∂ y4
y4 =
w̄m

x1 =
2w̄ f

x2 =
2w̄ f

x3 =
2w̄ f

x4 = .
2w̄ f
Mean fitness in males and females, respectively, for X-linked
loci will be
w̄m = x1 m 1 + x2 m 2 + x3 m 3 + x4 m 4

⎛ ⎞
0 0
∂ x1 ∂ x1 ∂ x1 ∂ x1 ∂ x1 ∂ x1 When A2 is fixed (y3 , x3 = 1), the leading eigenvalue (associated
∂ x2 ∂ y2 ∂ x3 ∂ y3 ∂ x4 ∂ y4
⎜ ⎟ with A2 B2 haplotype) is
⎜0 ∂ y1 ∂ y1 ∂ y1
0 ⎟
⎜ 0 ∂ x2
0 ∂ x3
0 ∂ x4 ⎟
⎜ ⎟
⎜0 0
∂ x2 ∂ x2
0 0
∂ x2 ∂ x2 ⎟
λB =
m 32
+
f 32
.
⎜ ∂ x2 ∂ y2 ∂ x4 ∂ y4 ⎟
⎜ ⎟ 2w̄m 2w̄ f
⎜0 0
∂ y2
0 0 0 0 0 ⎟
⎜ ∂ x2 ⎟
J X |x1 =y1 =1 =⎜ ∂ x3 ∂ x3 ∂ x3 ∂ x3 ⎟ .
⎜0 0 0 0 ⎟ When A2 is polymorphic (equilibrium frequency p̂), exact
⎜ ∂ x3 ∂ y3 ∂ x4 ∂ y4 ⎟
⎜ ∂ y3 ⎟ eigenvalues are more complex. However, because both loci (A
⎜0 0 0 0 0 0 0 ⎟
⎜ ∂ x3 ⎟
⎜ ∂ x4 ∂ x4 ⎟
and B) are cis-acting, we can approximate them by assuming
⎜0 0 0 0 0 0 ⎟
⎝ ∂ x4 ∂ y4 ⎠ tight linkage and ignoring squared recombinational terms (rm 2 ,
∂ y4
0 0 0 0 0 0 ∂ x4
0 rf 2 , rm rf ≈ 0). This yields a leading eigenvalue:
Because both matrices follow the block triangular form, their (1 − p̂) f 22C (1 − r f ) + p̂ f 32 (1 − p̂)m 22C (1 − rm ) + p̂m 32
λB ≈ +
2w̄ f 2w̄m

eigenvalues can be determined from each submatrix along the
For X-linked inheritance, the relevant matrix is:
⎛ m3
⎞
0 w̄m
0 0
⎜ x1 f 12 +x3 f 22R (1−r f ) y1 f 12 +y3 f 22R (1−r f ) ⎟
⎜ r f y1 f 22C r f x1 f 22C ⎟
⎜ 2w̄ f 2w̄ f 2w̄ f 2w̄ f ⎟
J X |[B1 ]=1 =⎜ x1 f 22C (1−r f )+x3 f 32 ⎟
⎜ r f x3 f 22R r f y3 f 22R y1 f 22C (1−r f )+y3 f 32 ⎟
⎝ 2w̄ f 2w̄ f 2w̄ f 2w̄ f ⎠
m4
0 0 w̄m
0
diagonal (Otto and Day 2007). Excluding the first 2 × 2 matrix with characteristic polynomial:
of zeros, the leading eigenvalues of the remaining three sub- det(J X − λI )
matrices are each candidates for the leading eigenvalue of the
m 3 (1 − p̂) f 12 + p̂ f 22R (1 − r f )
overall matrix. These three candidate eigenvalues are reported in = λ+ − λ2
w̄m 2w̄ f
Table 2, for each mode of inheritance.
m 4 (1 − p̂) f 22C (1 − r f ) + p̂ f 32
× λ+ −λ 2
w̄m 2w̄ f
Appendix 3: Stability with A2
rf
2
m3

m4

Present and B2 Absent − p̂(1 − p̂) f 22R f 22C
2w̄ f
λ+
w̄m
λ+
w̄m
.
(1) TIGHT LINKAGE BETWEEN LOCUS A AND B
When A2 is fixed, the leading eigenvalue is
To evaluate stability of the modifier locus under the sequential
invasion model, we recalculated the Jacobian matrix with B1 fixed f 32 w̄ f m 4
λB = 1+ 1+8 .
and A2 at arbitrary frequency x3 , which represents the equilibrium 4w̄ f w̄m f 32
for A2 when B1 is fixed. The analysis yields an 8 × 8 block
When A2 is polymorphic (again, assuming rf 2 ≈ 0)
triangular matrix, with one 4 × 4 matrix representing the resident
allele B1 and another 4 × 4 representing the nonresident allele (1 − p̂) f 22C (1 − r f ) + p̂ f 32
λB ≈
B2 . By definition, all eigenvalues of the resident submatrix be less 4w̄ f
than one (the system is stable with B1 fixed). Thus, we focus on w̄ f m 4
× 1+ 1+8
the second submatrix. w̄m [(1 − p̂) f 22C (1 − r f ) + p̂ f 32 ]
For autosomal inheritance, the relevant matrix is:
⎛ x1 m 12 +x3 m 22R (1−rm ) y1 m 12 +y3 m 22R (1−rm ) rm x1 m 22C rm y1 m 22C ⎞
2w̄m 2w̄m 2w̄m 2w̄m
⎜ x1 f 12 +x3 f 22R (1−r f ) y1 f 12 +y3 f 22R (1−r f ) ⎟
⎜ r f x1 f 22C r f y1 f 22C ⎟
⎜ 2w̄ f 2w̄ f 2w̄ f 2w̄ f ⎟
J A |[B1 ]=1 =⎜ ⎟.
⎜ rm x3 m 22R rm y3 m 22R x1 m 22C (1−rm )+x3 m 32 y1 m 22C (1−rm )+y3 m 32 ⎟
⎝ 2w̄m 2w̄m 2w̄m 2w̄m ⎠
r f x3 f 22R r f y3 f 22R x1 f 22C (1−r f )+x3 f 32 y1 f 22C (1−r f )+y3 f 32
2w̄ f 2w̄ f 2w̄ f 2w̄ f
Substituting p̂ for the equilibrium frequency of A2 (x3 ≈ y3

under weak selection), the characteristic polynomial is

2 (1 − p̂) f 22C (1 − r f ) + p̂ f 32 (1 − p̂)m 22C (1 − rm ) + p̂m 32 (1 − p̂)m 12 + p̂m 22R (1 − rm )
det(J A − λI ) = λ + −λ
2w̄ f 2w̄m 2w̄m

(1 − p̂) f 12 + p̂ f 22R (1 − r f ) rm (1 − p̂)m 22C r f (1 − p̂) f 22C rm p̂m 22R r f p̂ f 22R
+ − λ − λ2 + + .
2w̄ f 2w̄m 2w̄ f 2w̄m 2w̄ f

(2) LOOSE LINKAGE BETWEEN LOCUS A AND B particularly given our precondition that s(1 − hd ) > thd —the
For strong recombination relative to selection (ravg selection), minimum requirement for A2 to invade (eq. 4a). For X-linked
and small selection coefficients, we can approximate the eigen- inheritance and male-specific selection, B2 is always favored when
value associated with a B2 mutation, assuming linkage equilib- A2 invades because A2 always becomes fixed. For female-specific
rium between A and B (for similar analyses, see Otto and Bourguet selection, invasion of B2 will occur as long as s(1 − hd ) > rf ,
1999; Takahasi and Tajima 2005). When both loci are autosomal: which is likely to occur given our precondition for A2 invasion:
s(1 − hd ) > thd (eq. 4a)
(1 − p̂)2 m 12 + p̂(1 − p̂)(m 22C + m 22R ) + p̂ 2 m 32
λB =
2w̄m
(1 − p̂)2 f 12 + p̂(1 − p̂)( f 22C + f 22R ) + p̂ 2 f 32 (3). ANTAGONISTIC PLEIOTROPY; A2 POLYMORPHIC;
+ . LOOSE LINKAGE BETWEEN A AND B
2w̄ f
In contrast to the tight linkage scenario, where invasion is primar-
For two loosely linked X-linked loci ily determined by the A2 B2 interaction (relative fitness f 22C and

m22C ), a rare B2 allele will experience selection in a variety of ge-
(1 − p̂)m 3 + p̂m 4
λB = netic backgrounds. If A2 is rare, and B2 is sufficiently deleterious
3w̄m
(1 − p̂)2 f 12 + p̂(1 − p̂)( f 22C + f 22R ) + p̂ 2 f 32 in an A1 genetic background (e.g., f12 < f 11 and/or f 22R < f 21 ),
+2 . invasion may be impossible.
3w̄ f
Under the most permissive scenario, where B2 is neutral on an
A1 background and A2 B2 is beneficial (i.e, the parameterization of
Table Sl), the invasion of B2 will always be favored as long as A2 is
Appendix 4: Conditions Favoring present within the population. For the purpose of illustrating this
Invasion of B2 , Following result, we use the parameterization from Table S1, assuming male-
specific selection and autosomal linkage for both loci. Invasion
Invasion of A2 of B2 is favored when:
(1) ANTAGONISTIC PLEIOTROPY; A2 FIXED
For an autosomal locus and A2 fixed, the leading eigenvalue p̂(1 − p̂)h d t + p̂ 2 (1 − h d )t
0<
(which is associated with the A2 B2 haplotype) is w̄m
1 1 − hd t which simply requires that A2 be present within the population.
λB = +
2 2(1 − t)
which is greater than one as long as hd < 1. Since this is always (4). SEXUAL ANTAGONISM; A2 FIXED
true, B2 is always favored. For A2 fixed at an X-linked locus, Following fixation of a male-beneficial allele, a modifier of sex-
invasion is favored when: limitation will invade when:
2w̄ f − f 32 m4 1 − hd s f
< 1<
f 32 w̄m 1 − sf
When selection occurs in males, this reduces to t > 0, which which simply requires that hd < 1, which is always true. The same
is always true. When selection occurs in females, it reduces to results apply to modifiers following the invasion of a female-
hd < 1, which is always true. beneficial, autosomal mutation.
For X-linked inheritance, invasion of a modifier is favored
(2). ANTAGONISTIC PLEIOTROPY; A2 POLYMORPHIC; when:
TIGHT LINKAGE BETWEEN A AND B 2w̄ f − f 32 m4
<
For autosomal linkage, B2 is favored under male- and female- f 32 w̄m
specific selection, respectively, when:
which requires that hd < 1 following fixation of a male-beneficial
1 − r f (1 − p̂) (1 − p̂)(1 − h d s)(1 − rm ) + p̂(1 − h d t) allele and sm > 0 following fixation of a female-beneficial allele.
1< +
2 2w̄m Both will always be true.
1 − rm (1 − p̂) (1 − p̂)(1 − h d s)(1 − r f ) + p̂(1 − h d t) (5). SEXUAL ANTAGONISM; A2 POLYMORPHIC;

1< +
2 2w̄ f TIGHT LINKAGE BETWEEN A AND B
As A2 goes to fixation (p → 1), the previous results apply. Under autosomal inheritance, selection always favors the mod-
Alternatively, as p → 0, the minimal requirement for invasion is ifier when sm (1 − hd ) > rm + rf (following invasion of a
rm + rf < s(1 − hd ), which is fairly permissive for tight linkage, male-beneficial allele) and sf (1 − hd ) > rm + rf (following

invasion of a female-beneficial allele). These minimum criteria Thus, the probability that a B2 allele on an A1 -bearing chro-
are removed as the A2 equilibrium increases in frequency. mosome will escape to an A2 -bearing chromosome prior to
Under X-linked inheritance, and following invasion of a male loss from the population is:
beneficial allele, the modifier always invades when sm > 2rf ,
∞
t−1
with this condition relaxed as the equilibrium frequency of A2 1
Pr(escape) = 1− [1− p̂(rm +r f )/2]t−1 1−
increases. Following female-beneficial invasion, a modifier will t=1
N A (1− p̂)
invade as long as sf (1 − hd ) > rf , with this restriction removed 1 N A p̂(1 − p̂)(rm + r f )
× ≈ ,
for higher frequencies of A2 . N A (1 − p̂) N A p̂(1 − p̂)(rm + r f ) + 2
p̂(r +r )
(6). SEXUAL ANTAGONISM; A2 POLYMORPHIC; where the approximation assumes that 2N Am(1− fp̂) ≈ 0. For the X-
LOOSE LINKAGE BETWEEN A AND B linked scenario, the probability of escape is:
As it was for the case of antagonistic pleiotropy (see Appendix 4, ∞
t−1
1

section 3), invasion of a modifier with loose linkage between loci Pr(escape) = 1 − [1 − 2 p̂r f /3]t−1
1−
t=1
N X (1 − p̂)
depends on two factors: the degree to which B2 is deleterious in an
1 2N X p̂(1 − p̂)r f
A1 genetic background; and the frequency of A2 in the population. × ≈ .
N X (1 − p̂) 2N X p̂(1 − p̂)r f + 3
If A1 B2 is neutral, invasion will always be favored when p̂ > 0.
The waiting time until A2 B2 invades, when B2 is initially
linked to an A1 allele, is
Appendix 5: Waiting Times for 1
Modifier Invasion, Following Taut (inv.|A1 B2 ) =
2N A (1 − p̂)v A Pr(escape)(λ B − 1)
Invasion of A2 and Convergence =
N A p̂(1 − p̂)(rm + r f ) + 2
to Equilibrium 2N A p̂(1 − p̂)2 v A (rm + r f )(λ B − 1)
2
When A1 and A2 are polymorphic, there are two invasion routes

under autosomal inheritance, and
for B2
1
(1) If a B2 mutation arises on an A2 chromosome (with prob- TX (inv.|A1 B2 ) =
2N X (1 − p̂)v X Pr(escape)(λ B − 1)
ability p̂), the A2 B2 haplotype will invade with probability
2N X p̂(1 − p̂)r f + 3
2(λB − 1). The mean waiting time for such an event on the =
4N X2p̂(1 − p̂)2 v X r f (λ B − 1)
autosomes and X, respectively, is
under X-linked inheritance.
1 Taking both routes into account, the total waiting time until
Taut (inv.|A2 B2 ) =
2N A p̂v A (λ B − 1) B2 invades is
and
1
T (B2 ) = .
TX (inv.|A2 B2 ) =
1
, 1/T (inv.|A1 B2 ) + 1/T (inv.|A2 B2 )
2N X p̂v X (λ B − 1)
As selection against A1 B2 haplotypes increases (i.e., for A1 B2
where vA and vX refer to the chromosome-specific rate of mutation deleterious), or the A2 equilibrium approaches a relatively high
from B1 to B2 , and λB is the leading eigenvalue for the B locus, frequency within the population ( p̂ → 1), virtually all invasion
assuming A2 at frequency p̂. events proceed from an initial A2 B2 genetic background. This
(2) If B2 initially arises with A1 (with probability 1 − p̂), recom- suggests that the mean waiting time until B2 invasion will fall
bination with an A2 B1 haplotype can produce A2 B2 , which within the range is
once again invades with probability 2(λB − 1). The prob- 1
ability of such an event strongly depends on the marginal < T (B2 ) < T (inv.|A2 B2 ).
1/T (inv.|A1 B2 ) + 1/T (inv.|A2 B2 )
fitness of an A1 B2 haplotype within a population otherwise
fixed for B1 . In the most permissive case, A1 B2 is effectively Substituting the autosomal and X-linked terms, the range for
neutral (i.e., A1 B1 and A1 B2 have the same marginal fitness, each chromosome will be
as parameterized in Tables S1 and S2). Without recombi- N A p̂(1 − p̂)(rm + r f ) + 2
nation, the number of individuals inheriting a neutral A1 B2 2N A p̂v A (λ B − 1)[N A (1 − p̂)(rm + r f ) + 2]
haplotype destined to be lost from the population will be ge- 1
< Taut (B2 ) <
ometric with mean N A (1 − p̂) (Garcı́a-Dorado et al. 2003). 2N A p̂v A (λ B − 1)

and When A2 is rare and A1 B2 is neutral, the waiting time ap-

2N X p̂(1 − p̂)r f + 3 proaches the term on the left. As the equilibrium frequency of
2N X p̂v X (λ B − 1)[2N X (1 − p̂)r f + 3] A2 and the deleterious effect of A1 B2 increase, the waiting time
1 approaches the term on the right.
< TX (B2 ) < .
2N X p̂v X (λ B − 1)
Supporting Information
The following supporting information is available for this article:
Figure S1. Invasion probabilities for sexually antagonistic alleles.

Figure S2. Fixation probabilities for haplotypes with sex-specific benefits.
Figure S3. Antagonistic pleiotropy, modifier model.

Figure S4. Sexual antagonism, modifier model.
Figure S5. Double-mutant haplotype invasion model with no recombination in males.
Figure S6. Double-mutant haplotype invasion model with equal recombination in males and females.
Table S1. Full fitness parameterization for the two-locus pleiotropy modifier model, with B2 neutral on an A1 genetic background.
Table S2. Full fitness parameterization for the two-locus sexual antagonism modifier model, with B2 neutral on an A1 genetic
background.
Supporting Information may be found in the online version of this article.
Please note: Wiley-Blackwell is not responsible for the content or functionality of any supporting information supplied by the
authors. Any queries (other than missing material) should be directed to the corresponding author for the article.

Evolut 3417

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Evolut 3417

Uploaded by

Copyright:

Available Formats

O R I G I NA L A RT I C L E

SEX LINKAGE, SEX-SPECIFIC SELECTION, AND

Downloaded from https://academic.oup.com/evolut/article/64/12/3417/6853464 by guest on 17 March 2024

Received March 21, 2010

3417 Evolution 64-12: 3417–3442

Downloaded from https://academic.oup.com/evolut/article/64/12/3417/6853464 by guest on 17 March 2024

3418 EVOLUTION DECEMBER 2010

Table 1. Genotypes and corresponding sex-specific fitness parameters.

Female fitness: f 11 f 21 f 12 f 22C , f 22R f 31 f 13 f 32 f 23 f 33

Downloaded from https://academic.oup.com/evolut/article/64/12/3417/6853464 by guest on 17 March 2024

EVOLUTION DECEMBER 2010 3419

Locus A Locus B Both loci (A and B interacting)

Downloaded from https://academic.oup.com/evolut/article/64/12/3417/6853464 by guest on 17 March 2024

3420 EVOLUTION DECEMBER 2010

w(A2 A2 ) = exp{−c(x̂ − x1 − 2x)2 }. The dominance coeffi-

1 − exp {cx(2x̂ − 2x1 − x)}

Downloaded from https://academic.oup.com/evolut/article/64/12/3417/6853464 by guest on 17 March 2024

EVOLUTION DECEMBER 2010 3421

Downloaded from https://academic.oup.com/evolut/article/64/12/3417/6853464 by guest on 17 March 2024

3422 EVOLUTION DECEMBER 2010

The opportunity for each of these evolutionary transitions can and

Downloaded from https://academic.oup.com/evolut/article/64/12/3417/6853464 by guest on 17 March 2024

EVOLUTION DECEMBER 2010 3423

Table 3. Fitness parameterization under the antagonistic pleiotropy model.

Female-specific fitness effects

Downloaded from https://academic.oup.com/evolut/article/64/12/3417/6853464 by guest on 17 March 2024

1 An analysis of these eigenvalues indicates that, given invasion

3424 EVOLUTION DECEMBER 2010

mutation and recombination, and the strength of selection in favor

under autosomal inheritance, and:

Downloaded from https://academic.oup.com/evolut/article/64/12/3417/6853464 by guest on 17 March 2024

under X-linked inheritance (where NA and NX refer to the autoso-

EVOLUTION DECEMBER 2010 3425

Table 4. Fitness parameterization under the sexual antagonism model.

Selection for expression divergence in females

Downloaded from https://academic.oup.com/evolut/article/64/12/3417/6853464 by guest on 17 March 2024

3426 EVOLUTION DECEMBER 2010

Downloaded from https://academic.oup.com/evolut/article/64/12/3417/6853464 by guest on 17 March 2024

EVOLUTION DECEMBER 2010 3427

Downloaded from https://academic.oup.com/evolut/article/64/12/3417/6853464 by guest on 17 March 2024

3428 EVOLUTION DECEMBER 2010

Downloaded from https://academic.oup.com/evolut/article/64/12/3417/6853464 by guest on 17 March 2024

EVOLUTION DECEMBER 2010 3429

Downloaded from https://academic.oup.com/evolut/article/64/12/3417/6853464 by guest on 17 March 2024

3430 EVOLUTION DECEMBER 2010

Downloaded from https://academic.oup.com/evolut/article/64/12/3417/6853464 by guest on 17 March 2024

EVOLUTION DECEMBER 2010 3431

Downloaded from https://academic.oup.com/evolut/article/64/12/3417/6853464 by guest on 17 March 2024

3432 EVOLUTION DECEMBER 2010

Downloaded from https://academic.oup.com/evolut/article/64/12/3417/6853464 by guest on 17 March 2024

EVOLUTION DECEMBER 2010 3433

Downloaded from https://academic.oup.com/evolut/article/64/12/3417/6853464 by guest on 17 March 2024

3434 EVOLUTION DECEMBER 2010

Downloaded from https://academic.oup.com/evolut/article/64/12/3417/6853464 by guest on 17 March 2024

EVOLUTION DECEMBER 2010 3435

Downloaded from https://academic.oup.com/evolut/article/64/12/3417/6853464 by guest on 17 March 2024

3436 EVOLUTION DECEMBER 2010

Downloaded from https://academic.oup.com/evolut/article/64/12/3417/6853464 by guest on 17 March 2024

2x1 y1 m 11 + (x1 y2 + x2 y1 )m 12 + (x1 y3 + x3 y1 )m 21 + (1 − rm )m 22C (x1 y4 + x4 y1 ) + rm m 22R (x2 y3 + x3 y2 )

EVOLUTION DECEMBER 2010 3437

2x2 y2 f 13 + (x1 y2 + x2 y1 ) f 12 + (x2 y4 + x4 y2 ) f 23 + (1 − r f ) f 22R (x2 y3 + x3 y2 ) + r f f 22C (x4 y1 + x1 y4 )

2x3 y3 f 31 + (x1 y3 + x3 y1 ) f 21 + (x3 y4 + x4 y3 ) f 32 + (1 − r f ) f 22R (x2 y3 + x3 y2 ) + r f f 22C (x4 y1 + x1 y4 )

2x4 y4 f 33 + (x2 y4 + x4 y2 ) f 23 + (x4 y3 + x3 y4 ) f 32 + (1 − r f ) f 22C (x4 y1 + x1 y4 ) + r f f 22R (x2 y3 + x3 y2 )