ArticlePDF Available

Lenormand T.. Gene flow and the limit to natural selection. Trends Ecol Evol 17: 183-189

Authors:

Abstract and Figures

In general, individuals who survive to reproduce have genotypes that work relatively well under local conditions. Migrating or dispersing offspring elsewhere is likely to decrease an individual's or its offspring's fitness, not to mention the intrinsic costs and risks of dispersal. Gene flow into a population can counteract gene frequency changes because of selection, imposing a limit on local adaptation. In addition, the migrant flow tends to be higher from densely populated to sparsely populated areas. Thus, although the potential for adaptation might be greatest in poor and sparsely populated environments, gene flow will counteract selection more strongly in such populations. Recent papers, both theoretical and empirical, have clarified the important role of migration in evolution, affecting spatial patterns, species ranges and adaptation to the environment; in particular, by emphasizing the crucial interaction between evolutionary and demographic processes.
Content may be subject to copyright.
TRENDS in Ecology & Evolution
Vol.17 No.4 April 2002
http://tree.trends.com 0169-5347/02/$ – see front matter © 2002 Elsevier Science Ltd. All rights reserved. PII: S0169-5347(01)02432-6
183
Review
Thomas Lenormand
CEFE – CNRS, 1919 route
de Mende, 34293
Montpellier cedex 05,
France.
e-mail:
thomas.lenormand@
cefe.cnrs-mop.fr
Under natural selection, individuals tend to adapt
to their local environmental conditions, resulting
in a pattern of LOCAL ADAPTATION (see Glossary). Local
adaptation can occur if the direction of selection
changes for an allele among habitats (antagonistic
environmental effect), but it might also occur if
the intensity of selection at several loci that are
maintained as polymorphic by recurrent mutations
covaries negatively among habitats. These two
possibilities have been clearly identified in the related
context of the evolution of senescence but have not
have been fully appreciated in empirical and
theoretical studies of local adaptation [1,2].
The interaction between directional selection and
geneflow
When an allele with antagonistic environmental
effects is maintained at a MIGRATIONSELECTION
EQUILIBRIUM, gene flow changes allele frequencies in
a direction opposite to natural selection, and each
population is suboptimally adapted: that is, there is a
MIGRATION LOAD. For small amounts of migration, this
load per locus approximately equals the migration
rate. If the migration rate is large compared with
selection, the polymorphism is lost. When such GENE
SWAMPING occurs, the alleles with the best average
reproductive success across all populations tend to
become fixed; therefore there is no local adaptation
In general, individuals who survive to reproduce have genotypes that work
relatively well under local conditions.Migrating or dispersing offspring
elsewhere is likely to decrease an individual’s or its offspring’s fitness,not to
mention the intrinsic costs and risks of dispersal.Gene flow into a population
can counteract gene frequency changes because of selection, imposing a limit
on local adaptation. In addition, the migrant flow tends to be higher from
densely populated to sparsely populated areas. Thus, although the potential
for adaptation might be greatest in poor and sparsely populated environments,
gene flow will counteract selection more strongly in such populations. Recent
papers, both theoretical and empirical,have clarified the important role of
migration in evolution, affecting spatial patterns, species ranges and
adaptation to the environment; in particular, by emphasizing the crucial
interaction between evolutionary and demographic processes.
Gene flow and the limits to natural
selection
Thomas Lenormand
35 Giller, P.S. (1996) The diversity of soil
communities, the ‘poor man’s tropical rainforest’.
Biodivers. Conserv. 5, 135–168
36 Anderson, J.M. (1978) A method to quantify soil-
microhabitat complexity and its application to a
study of soil animal species diversity. Soil Biol.
Biochem.10, 77–78
37 Hansen, R. and Coleman, D.C. (1998) Litter
complexity and composition are determinants of
the diversity and composition of oribatid mites
(Acari: Oribatida) in litterbags. Appl. Soil Ecol.
9,17–23
38 Sulkava, P. and Huhta, V. (1998) Habitat
patchiness affects decomposition and faunal
diversity: a microcosm experiment on forest floor.
Oecologia116, 390–396
39 O’Connell, T. and Bolger, T. (1998) Intraspecific
aggregation, ‘probability niches’ and the diversity
of soil microarthropod assemblages. Appl. Soil
Ecol.9, 63–67
40 Bengtsson, G. (1997) Dispersal, heterogeneity
and resistance: challenging soil quality
assessment.In Ecological Risk Assessment of
Contaminants in Soil (Van Straalen, N. and
Lokke, H., eds), pp. 191–212, Chapman & Hall
41 Dighton, J. et al. (1997) The role of abiotic
factors, cultivation practices and soil fauna in
the dispersal of genetically modified
microorganisms in soil. Appl. Soil Ecol.
5,109–131
42 McIntyre, N.E. and Wiens, J.A. (1999) How does
habitat patch size affect animal movement? An
experiment with darkling beetles. Ecology
80,2261–2270
43 Felske, A. and Akkermans, A.D.L. (1998) Spatial
homogeneity of abundant bacterial 16S rRNA
molecules in grassland soils. Microb. Ecol. 36, 31–36
44 Wardle, D.A. (2002) Communities and Ecosystems:
Linking the Aboveground and Belowground
Components, Princeton University Press
45 Robertson, G.P. and Gross, K.L. (1994) Assessing the
heterogeneity of belowground resources: quantifying
pattern and scale.In Exploitation of Environmental
Heterogeneity by Plants: Ecophysiological Processes
Above- and Belowground (Caldwell, M.M. and
Pearcy, R.W., eds), pp. 237–253, Academic Press
46 Spain, A.V. and McIvor, J.G. (1988) The nature of
herbaceous vegetation associated with termitaria
in north-eastern Australia. J. Ecol. 76, 181–191
47 Hobbs, R.J. and Mooney, H.A. (1985) Community
and population dynamics of serpentine grassland
annuals in relation to gopher disturbance.
Oecologia67, 342–351
48 Packer, A. and Clay, K. (2000) Soil pathogens and
spatial patterns of seedling mortality in a
temperate tree. Nature404, 278–281
49 Olff, H. et al. (2000) Small-scale shifting mosaics
of two dominant grassland species: the possible
role of soil-borne pathogens. Oecologia125, 45–54
50 Brown, V.K. and Gange, A.C. (1989) Root
herbivory by insects depresses plant species
richness. Funct. Ecol. 3, 667–671
51 Van der Heijden, M. et al.(1998) Different
arbuscular mycorrhizal fungal species are
potential determinants of plant community
structure. Ecology79, 2082–2091
52 Hartnett, D.C. and Wilson, G.W.T. (1999)
Mycorrhizae influence plant community structure
and diversity in tall-grass prairie. Ecology
80,1187–1195
53 Campbell, B.D. et al. (1991) A trade-off between
scale and precision in resource foraging. Oecologia
87, 532–538
54 Hutchings, M.J. and De Kroon, H. (1994) The role
of morphological plasticity in resource
acquisition. Adv. Ecol. Res. 25, 160–238
55 Sulkava, P. et al. (2001) Influence of soil fauna and
habitat patchiness on plant (Betula pendula)
growth and carbon dynamics in a microcosm
experiment. Oecologia129, 133–138
56 Bonkowski, M. et al. (2000) Substrate
heterogeneity and microfauna in soil organic
‘hotspots’ as determinants of nitrogen capture and
growth of ryegrass. Appl. Soil Ecol. 14, 37–53
57 Hodge, A. et al. (2000) Spatial and physical
heterogeneity of N supply from soil does not
influence N capture by two grass species. Funct.
Ecol.14, 645–653
58 Korsaeth, A. et al. (2001) Modelling the
competition for nitrogen between plants and
microflora as a function of soil heterogeneity. Soil
Biol. Biochem. 33, 215–226
59 Robertson, G.P. et al. (1997) Soil resources,
microbial activity, and primary production
across an agricultural ecosystem. Ecol. Appl.
7,158–170
even though selection varies over space. Gene
swamping occurs even in simple models, where
population density is not affected by the amount of
selection or by the amount of dispersal (SOFT SELECTION
models). In the simplest possible model that exhibits
gene swamping, there are two habitats, soft selection,
one locus with antagonistic environmental effect and
three parameters: (1) the amount of migration m
(assuming that migration is symmetrical); (2) the
intensity of selection sin one habitat; and (3) the ratio
of selection coefficients between the two habitats
α
(Box 1). Gene swamping occurs if m/s>
α
/(1-
α
) for
α
chosen such that
α
<1 [3]. Thus, if the selection
coefficients do not perfectly balance between habitats
(
α
<1), gene swamping occurs in a large portion of the
parameter space.
This basic model can be elaborated introducing
an asymmetry in density or migration between the
two habitats; this tends to favor the allele that is
advantageous in the high density, low immigration
deme [3]. Another important illustration of gene
swamping comes from models of a continuous habitat
with a constant density of individuals. An allele that
is beneficial in a pocket of the environment might
decrease in frequency and be lost even where it has a
local selective advantage (Box 2). This type of model
shows that there are two crucial combinations of
parameters determining migration–selection
equilibria: one integrating the relative spatial scale
of migration and selection and the other combining
the different asymmetries among habitats.
Evidence for migration load and gene swamping
Mayr’s observation [4] that phenotypic divergence
is often correlated with the degree of isolation is
consistent with the idea that gene flow keeps
divergence in check by opposing the effect of natural
selection. This can also be explained by the action of
genetic drift or by variation of selection pressure with
distance. Evidence for a migration load comes from
more precise studies. The study of gene frequency
patterns across a sharp transition in the environment
(e.g. presence or absence of heavy metals in habitats of
Agrostistenuis and Anthoxantumodoratum[5]) often
results in the observation of smooth clines ([6] for
other examples). The width of these clines is typically
larger than the environmental transition and can
be used to estimate the opposing effects of selection
and gene flow (i.e. σ/s; Box 2). Transplantation
experiments can also provide a good estimate of the
opposing effects of dispersal on selection provided
that the scale of the transplantation is carried out at
the scale of dispersal (this requires an independent
measure of dispersal [7,8]). Further evidence for a
migration load is that the degree of local adaptation
(measured experimentally) correlates with the
degree of isolation of populations [9,10]. More
directly, temporal variation in the parameters
(gene flow and selection pressure) generates
variation in local adaptation that is consistent with
a migration–selection model [11,12].
Gene swamping is an extreme form of migration
load. However, it is much more difficult to
demonstrate because it requires one to show, in the
absence of a genetic polymorphism, that the genetic
potential for local adaptation exists. Gene swamping
can be inferred, however, if local adaptation occurs
only in sufficiently large or isolated environments. For
instance, Culexmosquitoes in Corsica do not exhibit
insecticide resistance in spite of the occurrence of the
selection pressure and the presence, at low frequency,
of resistance genes that have spread elsewhere. This
situation is best explained by gene swamping, because
the insecticide-treated area in this region is small
compared with the dispersal distance of Culex[13].
When the environment changes with distance
The maladaptive effects of dispersal depend strongly
on how selection varies over space. In models where
dispersal is a function of distance, one has to specify
how selection varies with distance. This specification
has an enormous impact on the theoretical
predictions. For instance, in a model of Felsenstein
[14] where the optimum for a quantitative trait varies
TRENDS in Ecology & Evolution
Vol.17 No.4 April 2002
http://tree.trends.com
184 Review
Figure I illustrates a model [a] where allele
A
has an
advantage
s
over allele
a
in habitat 1 and a disadvantage
αs
in habitat 2. A fraction
m
(the migration rate) of
individuals from each habitat is exchanged each
generation. With antagonistic environmental effect (
α
> 0),
A
is favorable in one habitat and
a
in the other. If migration
is strong enough relative to selection (
m
/
s
larger than a
critical value), the allele with the largest fitness averaged
over both habitats will tend to become fixed (e.g. it will be
A
if 0 <
α
< 1 and
s
>0). The critical value for
m
/
s
is
α
/ |1-
α
|,
which increases when the two alleles are on overall
equally fit (
α
close to 1). If
m
/
s
is less than this critical value,
a polymorphism is maintained by a balance between
migration and selection. If
m
/
s
is greater than this critical
value, the best overall allele ‘swamp’ the other.
If
α
is negative then
A
is either deleterious or
advantageous in both habitats, and recurrent mutation
is necessary to maintain a polymorphism. I considered
here only the interaction between allelic effect and the
environment. However, fitness can also depend on the
interaction between alleles at the same or different loci.
For instance, if selection occurs on diploids, marginal
over- or underdominance can produce other types of
stable polymorphism among habitats.
Reference
a Bulmer, M.G. (1972) Multiple niche polymorphism.
Am.Nat.106, 254–257
Box 1.Migration–selection model with two
demes and two alleles
Habitat 1 Habitat 2
m
1+
s
1
1
s
1
α
A
a
Relative
fitness
of allele
Fig. I
linearly along a single dimension (with constant
and equal densities everywhere and homogeneous
dispersal), the population mean of the trait matches
perfectly the optimum regardless of the amount of
migration. This extreme result is obtained because,
in any location, the genetic and phenotypic change
caused by migrants from the right and the left cancel
each other out exactly. Perfect matching is not
observed, however, if the environment is finite or if
the gradient is nonlinear, or if the underlying genes
do not contribute additively to the phenotype. The
repeated occurrence of wide genetic clines on different
continents has been used to infer that selection
pressures vary with distance (e.g. continental body
size clines in Drosophilasubobscura [15]). The
migration load occurring within these clines has not
been estimated but can be substantial because of
linkage disequilibria (i.e. when the environmental
gradient and dispersal are strong) or edge effects
(peripheral population or populations at the edge
of wide clines are likely to exhibit maladaptation
because of a relatively low density [16]).
The effect of genetic parameters, life cycle and mating
system
Several other important biological parameters
influence the impact of migration. First, an allelic
effect might depend on the presence of other alleles at
the same (dominance) or different (epistasis) locus,
as well as on the environment. These two types of
interaction have a similar and potentially large effect
on migration–selection equilibria. For instance, the
range of parameters where gene swamping occurs
is much larger for alleles whose local advantage is
recessive [17], especially when segregation occurs
between migration and selection within the life cycle.
Marginal underdominance because of different
dominance coefficients among environments could
lead to migration–selection equilibria that are locally
stable but that are dependent on initial conditions. In
these cases, gene swamping could occur very rapidly
following a large perturbation in allele frequencies
(e.g. following a bottleneck).
Second, migration produces associations between
the selected alleles. Within-loci associations, often
measured by heterozygote deficit, are removed
efficiently by segregation at meiosis. However,
between-loci associations, often measured by linkage
disequilibria, are only partially removed by
recombination at meiosis. For alleles whose effects
covary positively over space, migration creates a local
excess of extreme genotypes in fitness (i.e. positive
linkage disequilibrium). Recombination tends to
reduce this disequilibrium and thus also the variance
in fitness. As a consequence, the efficacy of selection
is lower and the migration load higher with high
recombination rates [18–21], especially when
recombination occurs between migration and selection
within the life cycle. Consequently, the migration load
will tend to be higher in models assuming many
TRENDS in Ecology & Evolution
Vol.17 No.4 April 2002
http://tree.trends.com
185
Review
Figure I shows the frequency pattern (dashed line) of an allele
A
that has an additive
fitness advantage of
s
in a 1D environmental ‘pocket’ of size 2
a
and is deleterious
elsewhere by an amount α
s
. The model allows the density (dotted line) outside
the environmental pocket to be τtimes the density inside the pocket. Gene flow is
measured by
σ
, the standard deviation of parent–offspring distance measured along
one dimension. At equilibrium, the maximum frequency of the beneficial allele within
the environmental pocket,
p
max, depends on two combinations of the parameters
(Fig. II) [a,b]. The first combination (y axis on Fig. II) measures the relative scales of
the spatial heterogeneity (
a
) and of the ‘characteristic length’ (
σ
/
s
), which weighs the
strength of selection relative to gene flow. The second combination (x axis on Fig. II)
is a measure of overall asymmetry between habitats (ratio of selection coefficients
and ratio of densities outside/inside the pocket). Figure II is read as a contour-plot
for
p
max. The blue area corresponds to the parameter space where gene swamping
occurs: allele
A
is absent (
p
max =0) from the geographical pocket in spite of its local
selective advantage. Gene swamping occurs if the relative scales parameter is small
and the asymmetry parameter large. The density ratio outside/inside the pocket (
τ
)
strongly magnifies the ratio of selection coefficients outside/inside the pocket (
α
):
the asymmetry parameter is
τ
2
α
1/2, which indicates that
α
plays the same role as the
fourth power of
τ
. However, the asymmetry between habitats has a limited impact on
local adaptation:
p
max is almost independent from the x axis when the asymmetry
parameter is >~3 (e.g. the curve
p
max =0 tends toward π/2). The parameter
combination leading to a maximum frequency
p
max between 0 and 0.5 is quite narrow
(green area): allele
A
is likely to be either absent or present at high frequency.
The above results for one locus can be extended to a quantitative trait under
stabilizing selection. Each underlying locus can be viewed as a locus with alleles with
a small effect that exhibits antagonistic pleiotropy over space. As a consequence, a
quantitative trait with additive variance
V
aunder stabilizing selection of intensity
s
cannot significantly track environmental change under a critical geographical scale,
which is proportional to a ‘characteristic length’ σ /(
V
a
s
) [c].
Reference
a Nagylaki, T. (1975) Conditions for the existence of clines. Genetics80, 595–615
b Nagylaki, T. (1978) Clines with asymmetric migration. Genetics88, 813–827
c Slatkin, M. (1978) Spatial patterns in the distributions of polygenic characters. J. Theor.
Biol. 70, 213–228
Box 2. Environmental pockets of adaptation
1
0
s
α
s
α
s
p
max
τ
1
0
Distance
aa
TRENDS in Ecology & Evolution
Frequency of
A
Density
10234567
π
p
max
= 0.99
π/2
/2
Asymmetry between habitats ( 2 1/2)
Relatives scale ( )
p
max
= 0.90
p
max
= 0.50
p
max
= 0.00
0.90<
p
max
< 0.99
0.50<
p
max
< 0.90
0.99<
p
max
< 1.00
p
max
= 0.00
TRENDS in Ecology & Evolution
a
2
s
σ
τα
Fig. I
Fig. II
unlinked loci of small effects (e.g. quantitative genetics
models [22,23]) than in single-locus models [17,24].
The mating system can also strongly modify the
impact of migration. When locally adapted genotypes
tend to have a greater mating success, the genetic
differences among populations are accentuated, which
generally reduces the migration load. For instance,
sexual selection might increase dramatically the
reproductive success of locally adapted males [25].
Assortative mating could also increase the mating
success of the most common genotypes, because they
encounter suitable mates more often. This mating
system will tend to favor the locally adapted genotypes,
because they are likely to be common where they are
favored. However, assortative mating can prevent local
adaptation whenever the locally adapted genotype
happens to be rare (e.g. when a mutation arises [25], in
situations close to gene swamping, or simply by chance
[26]). In addition, the mating system might not change
the mating success of individuals but only the genetic
correlations between mates, which thus affects the
variance in fitness. For instance, positive assortative
mating (e.g. because of an assortment trait, such as
flowering time [26], behavioral imprinting [27] or
selfing [28]) will, in general, increase the variance in
fitness and thus reduce the migration load. Overall,
any parameter that increases the efficacy of selection,
by increasing either the relative fitness of locally
adapted genotypes or the local genetic variance in
fitness, will reduce the migration load.
Relaxing soft selection hypotheses
When the assumptions of soft selection are relaxed
(i.e. under HARD SELECTION), the qualitative impact
of gene flow on adaptation is less straightforward;
introducing more realistic demography within local
adaptation models has interesting, but complicated,
consequences. For clarity, I separate the interaction
TRENDS in Ecology & Evolution
Vol.17 No.4 April 2002
http://tree.trends.com
186 Review
Box 3. Evolution out of a niche
( 1)/
αα
0.001 0.01 0.1 1
Migration rate
(b)
(a)
Soft selection:
A
Hard selection:
A
Soft selection:
A
Hard selection:
A
r
10
r
100
r
0.1
r
0.01
r
Soft selection:
A
Hard selection:
A
Soft selection:
A
Hard selection:
A
Soft selection:
A
Hard selection:
A
r
10
r
100
r
0.1
r
0.01
r
TRENDS in Ecology & Evolution
1
α
r
α
Minimum selection coefficient for the allele beneficial
in the sink to increase in frequency when rare
Table I Habitat 1 Habitat 2
Intrinsic rate of increase
a
rr
A
r
+
sr
-
αs
Here, I consider the fate of an allele
A
that increases the
intrinsic growth rate in habitat 1,
r
, by an amount
s
but
reduces intrinsic growth rate in habitat 2,
r
, by an amount
αs
(Table I).
The two habitats are connected by migration at rate
m
per individual. In the absence of
A
, only habitat 2 has a
positive growth rate (-
r
<0 and
r
>0) and is self sustaining
(the population in habitat 1 is maintained by the recurrent
immigration from habitat 2). For simplicity, regulation by
density occurs only in habitat 2. I compare this hard selection
model [a] with the soft selection model presented in Fig. I
in Box 1, where population sizes in habitats 1 and 2 are
considered constant and equal.
In Fig. I, the solid and dashed curves indicate the
minimum selection intensity (y axis) required for the allele
A
to increase when rare for a given migration (x axis) and a
given ratio of selection coefficients
α
under hard and soft
selection, respectively.
Figure Ia considers the case where
A
is more
advantageous in habitat 1 than disadvantageous in habitat 2
(α <1), whereas Fig. Ib considers the opposite case (α > 1).
When
A
is, on average, advantageous (Fig. Ia), it always
increases in frequency when rare in the soft selection case. It
might even fix if migration is high enough (Box 1). However,
this is no longer true with hard selection:
A
is lost in the
green area. With hard selection,
A
increases in frequency
only if it is beneficial enough to compensate for the negative
growth rate in habitat 1, especially when migration is so low
that only a small fraction of the total population is exposed
to selection in habitat 1. When
A
is, on average, deleterious
(Fig. Ib), it increases in frequency when rare in the soft
selection case only if migration is low enough (i.e. above
the dashed line, see Box 1). However, as in Fig. Ia, with hard
selection and low migration, allele
A
also has to be beneficial
enough to compensate for the negative growth rate in
habitat 1. In both Fig. Ia and b, the conditions under which
A
increases in frequency are more stringent under hard than
under soft selection (because
A
is the allele beneficial in the
bad habitat) and the difference between hard and soft
selection decreases when the intensity of selection (
s
) or
migration (
m
)increases.
References
a Kawecki, T.J. (2000) Adaptation to marginal habitats:
contrasting influence of the dispersal rate on the fate of alleles
with small and large effects. Proc. R. Soc. London B Biol. Sci.
267, 1315–1320
Fig. I
between mean fitness and density and the interaction
between migration and density, although both effects
occur simultaneously (Box 3).
Migration meltdown: linking density and mean fitness
When local density is a monotonically increasing
function of mean fitness, a positive feedback loop can
occur that magnifies the effect of gene swamping:
populations with a lower-than-average density
receive a proportionately higher-than-average influx
of maladapted alleles, their mean fitness decreases,
which reduces their density further, and so on. Thus,
asymmetries in gene flow can become more extreme
over time, which destabilizes migration–selection
equilibria. For example, differences in density
occurring by chance (e.g. following an environmental
perturbation) can arbitrarily cause one allele to
swamp another. This MIGRATION MELTDOWN is similar
to the feedback loop occurring in mutation meltdown
models [2], and indeed the two might work in concert.
Migration meltdown might also be enhanced by a
mating system that confers a mating advantage to
the most common genotype [25].
Migration meltdown has been proposed as an
explanation for constraints on species range [22]. In
this model, a quantitative trait whose optimum
changes with distance is considered. Because density
decreases when mean fitness decreases, a migration
meltdown occurs at the species boundary whenever
migration is too large. The same process occurs in
models considering the evolution of ecological
specialization. An accidental change in population size
or habitat availability can be followed by a migration
meltdown, resulting in a niche switch or reduction [23].
Direct evidence for a migration meltdown is almost
entirely absent and probably very difficult to obtain.
Studies of the butterfly Euphydryaseditha in
California [29] document that a niche switch between
the two main host plants of this butterfly occurred
after an unusual summer frost, which killed all the
larvae on the preferred host plant Collinsia. In the
subsequent years, the butterflies consistently
reproduced on the other host plant Pedicularisin
spite of normal weather conditions. The reason for
this switch could be because recolonization of
Collinsiais difficult (the life cycle on Pedicularis
occurs later in the season, which creates a temporal
barrier to migration) but it might also be due to
genetic changes that were induced by the frost [23].
Migration rescue: the movement of individuals
When gene flow results directly from a movement of
individuals (e.g. seed dispersal but not pollen dispersal
when pollen is not limiting), asymmetric gene flow can
have demographic consequences. With all individuals
having the same probability of migrating, habitats of
lower-than-average density will have a net increase in
density after dispersal and vice versa. By decreasing
density differences among habitats, the migration of
individuals opposes the demographic effects of a
migration meltdown [23]. Net immigration can
compensate for the reduction in density caused by
maladaptation to local conditions. Whether the
demographic or genetic effect of migration dominates
depends ultimately on the strength of the relationship
between density and adaptation, about which we often
know very little. It is this type of contradictory effect
that has to be balanced in agronomy or conservation
biology when some populations are artificially
supplemented with maladapted individuals for a
demographic reason [30,31].
At one extreme, dispersal can sustain some
populations at non-zero densities that would
otherwise go extinct [32]. This situation could have
profound evolutionary consequences, because it
might allow the niche of a species to evolve. Dispersal
brings individuals into new environments and
therefore makes adaptation to these new conditions
possible. This effect, known as ‘evolution-out-of-a-
niche’ [24,33,34], is particularly important for alleles
that confer an advantage in the new habitat that
is not great enough for the population to be self-
sustaining in that habitat without immigration.
This kind of allele can spread only with very high
rates of dispersal (Box 3).
A genetic rescue effect can also occur because
dispersal increases genetic variance, which is
necessary for adaptation but which might otherwise
be absent in small peripheral populations. This effect
has been quantified by Gomulkiewicz et al. [34], who
showed that local adaptation can proceed faster with
intermediate migration rates. Dispersal can also
reduce the deleterious effects of inbreeding acting in
small peripheral populations [35], although this effect
might be weak compared with the migration load
owing to local adaptation [28].
The interaction of gene flow and disruptive selection
A migration load can also be generated by interactions
among alleles without any effect of the environment:
immigrant alleles might simply not work well with
combinations of local alleles. This occurs when there
are alternative adaptive peaks within a single habitat.
Such a disruptive selection pattern can be caused by
underdominance, strong epistasis or, more generally,
selection against rare genotypes [36]. The transition
between directional and disruptive selection is
gradual and depends on the relative depth of the
‘adaptive valley’ and the altitude difference between
the peaks. Moreover, both types of selection can lead to
the same pattern (e.g. the same migration–selection
equilibria [37]) and have been often grouped and
referred to as outbreeding depression or hybrid
breakdown [38]. Their relative importance has largely
been evaluated (and debated) in the context of hybrid
zones using crosses and transplants [39]. However,
decomposing the factors causing outbreeding
depression within species could be achieved by the
same method, which would be useful for evaluating
the effect of gene flow on adaptation.
TRENDS in Ecology & Evolution
Vol.17 No.4 April 2002
http://tree.trends.com
187
Review
One of the most controversial debates in
evolutionary biology has been the role of gene flow
in adaptation under disruptive selection between at
least two fitness peaks of different height (a process
referred to as the shifting balance or SBP [40]). Gene
flow plays a crucial role in shifting balance models; it
has to be both low enough for a peak shift to occur (so
that drift enables alleles to cross the ‘valley’) and high
enough for a high-fitness peak to spread [41]. Thus,
adaptation by the SBP, if it occurs, requires rather
intermediate levels of gene flow. Demographic effects
of migration might help or hinder the SBP. If deme
density increases with mean fitness, demes near
the high fitness peak might contribute more to the
migrant pool [42] and spread more rapidly. However,
a similar process to the migration meltdown, the
HYBRID SINK EFFECT [43], can decrease local gene flow
in demes at the boundary between populations on
different peaks (because such demes will tend to fall
in the fitness valley between the peaks). A local
reduction in gene flow (a barrier) can dramatically
slow the spread of an adaptive peak. The presence of
a barrier increases the effect of disruptive selection
because migrants are rarer on the other side of the
barrier and are thus more heavily selected against.
Barriers can therefore prevent the spread of a high
fitness peak [43]. Overall, gene flow can severely limit
adaptation when there is a significant amount of
disruptive selection whenever migration rates are too
high, too low or too irregular over space. That hybrid
zones often coincide in space with natural barriers is
consistent with this last effect [44].
Indirect effect of the migration load
The fitness load because of migration can be very high
in a heterogeneous environment, which creates a
strong indirect selection pressure to suppress it.
The migration load can be alleviated by a genetic
amelioration process [45], whereby the effects of genes
are modified directly (e.g. by increased dominance [46]
or decreased deleterious pleiotropic effects [47]).
This migration load also generates a weak indirect
selection pressure that favors any trait that reduces
gene flow between habitats (e.g. increased assortative
mating, increased habitat choice, reduced movement
rate, etc.), a mechanism which works in the same
way as reinforcement. However, these traits can be
classified into two categories depending on whether
they have to differ between habitats to reduce gene
flow (one- versus two-allele mechanisms [48]). For
instance, habitat preference can be caused by habitat-
specific alleles (e.g. an individual might prefer habitat
A because it carries allele A(a two-allele mechanism))
or by alleles that are not habitat specific (e.g. an allele
which causes individuals to prefer the habitat where
they were born (a one-allele mechanism)). The
indirect selection pressures caused by the migration
load tend to be weak and thus might, in general,
be insufficient to oppose gene flow and produce
divergence among populations of a reinforcement
trait [26], although there is evidence that it can
occur [49]. As a consequence, one-allele mechanisms
(e.g. reduced dispersal [28,50], increased plasticity
[1,51,52], reduced recombination [20,21], imprinting-
based mating or habitat preferences) might be more
likely to evolve to reduce the migration load than are
two-allele mechanisms.
These different processes can occur only if there
is a migration load. However, most of them cannot
operate if there is gene swamping. For instance, the
amelioration process cannot occur if the polymorphism
at the selected loci is swamped by gene flow. In this
case, the rate of adaptation is determined by the rate
of appearance of beneficial mutations that have few
deleterious pleiotropic effects in the different habitats.
To some extent, gene flow might prevent evolution
by mutations of small effect; assuming that a small
beneficial effect in one place is more easily swamped by
deleterious effects elsewhere, spatial heterogeneity in
selection might therefore bias upward the average
effect of alleles that underlie adaptation [53].
Dispersal evolution
Dispersal, which is the ultimate source of migration
load, is selected against in a heterogeneous habitat.
Gene flow is therefore only a proximate factor limiting
adaptation, the ultimate factors being the selective
forces (or constraints) that favor dispersal. Dispersal
might simply be unavoidable, but it can also be
directly favored for different reasons. First, dispersal
can evolve because the environment varies in time
and space (to escape crowding [54] or as a risk-
spreading strategy [55]). Second, dispersal could
evolve as an altruistic trait to reduce kin competition
[56]. Third, dispersal could be favored to escape the
fitness depression caused by mating among relatives
when there are many deleterious recessive mutations
[57]. Fourth, gene flow could be favored because it
TRENDS in Ecology & Evolution
Vol.17 No.4 April 2002
http://tree.trends.com
188 Review
Gene swamping: loss of the genetic variance at a locus under selection because gene flow is too
high.
Sensustricto
, gene swamping cannot occur when genetic variance is maintained by
recurrent mutations (e.g. in quantitative genetic models). However, the concept can used
sensulato
to describe situations where there is no significant response to selection because
gene flow is too high. This occurs when the spatial scale of environmental heterogeneity is
below a ‘characteristic length’ [a] equal to σ /
s
(Box 2).
Hybrid sink effect: self-reinforcing process in which immigration causes mating between two
subspecies, which produces unfit hybrids, which decreases local density, which increases
immigration rate and so on.
Local adaptation: a better average performance of individuals born in the habitat in which the
measure is done, compared with the performance of immigrants.
Migration load: decrease in mean fitness of a population because of immigration. This occurs
because the phenotypic mean of the population is different from the local optimum value.
Migration meltdown: self-reinforcing process in which immigration brings locally maladapted
alleles, which decreases local density, which increases immigration rate and so on.
Migration–selection equilibrium: stable polymorphic equilibrium in which the frequency change
caused by migration cancels out the frequency change caused by selection.
Soft and hard selection: soft selection refers to models where population density is not affected
by the amount of selection or by the amount of dispersal; hard selection refers to any other cases.
Reference
a Slatkin, M. (1978) Spatial patterns in the distributions of polygenic characters. J. Theor.
Biol. 70, 213–228
Glossary
Acknowledgements
I thank P. David,
T. Guillemaud, P. Jarne,
M. Kirkpatrick, S. Otto,
O. Ronce and M. Whitlock
for insightful discussion
and useful comments.
This study was supported
by the Centre National de
la Recherche Scientifique
(CNRS) and French
Ministry of Research.
References
1 Whitlock, M.C. (1996) The red queen beats the
jack-of-all-trades: the limitations on the evolution
of phenotypic plasticity and niche breadth. Am.
Nat.148, S65–S77
2 Kawecki, T.J.et al. (1997) Mutational collapse of
fitness in marginal habitats and the evolution of
ecological specialisation. J. Evol. Biol. 10, 407–429
3 Bulmer, M.G. (1972) Multiple niche
polymorphism. Am. Nat. 106, 254–257
4 Mayr, E. (1963) Animal, Species and Evolution,
Harvard University Press
5 Jain, S.K. and Bradshaw, A.D. (1966)
Evolutionary divergence among adjacent plant
populations. I. The evidence and its theoretical
analysis. Heredity21, 407–441
6 May, R.M. et al. (1975) Gene frequency clines in
the presence of selection opposed by gene flow.
Am. Nat. 109, 659–676
7 Via, S. (1991) The genetic structure of host plant
adaptation in a spatial patchwork – demographic
variability among reciprocally transplanted pea
aphid clones. Evolution45, 827–852
8 Burt, A. (1995) The evolution of fitness. Evolution
49, 1–8
9 Stearns, S.C. and Sage, R.D. (1980) Maladaptation
in marginal population of the mosquito fish,
Gambusia affinis. Evolution34, 65–75
10 Storfer, A. et al. (1999) Adaptive coloration and
gene flow as a constraint to local adaptation in the
streamside salamander, Ambystoma barbouri.
Evolution53, 889–898
11 Riechert, S. (1993) Investigation of potential gene
flow limitation of behavioral adaptation in an
aridlands spider. Behav. Ecol. Sociobiol.
32,355–363
12 Lenormand, T.et al. (1999) Tracking the evolution
of insecticide resistance in the mosquito Culex
pipiens. Nature400, 861–864
13 Raymond, M. and Marquine, M. (1994) Evolution of
insecticide resistance in Culex pipiens populations:
the Corsican paradox. J. Evol. Biol. 7, 315–337
14 Felsenstein, J. (1977) Multivariate normal
genetic model with a finite number of loci. In
Proceedings of the International Conference on
Quantitative Genetics (Pollak, E. et al., eds),
pp. 227–246, Iowa State University Press
15 Huey, R.B. et al. (2000) Rapid evolution of a
geographic cline in size in an introduced fly.
Science287, 308–309
16 Garcia-Ramos, G. and Kirkpatrick, M. (1997)
Genetic models of adaptation and gene flow in
peripheral populations. Evolution51, 21–28
17 Nagylaki, T. (1975) Conditions for the existence of
clines. Genetics80, 595–615
18 Slatkin, M. (1975) Gene flow and selection in a
two locus system. Genetics81, 787–802
19 Lenormand, T. and Raymond, M. (1998)
Resistance management: the stable zone strategy.
Proc. R. Soc. London B Biol. Sci. 265, 1985–1990
20 Pylkov, K.V.et al. (1998) Migration versus
mutation in the evolution of recombination under
multilocus selection. Genet. Res. 71, 247–256
21 Lenormand, T. and Otto, S.P. (2000) The evolution
of recombination in a heterogeneous
environment. Genetics156, 423–38
22 Kirkpatrick, M. and Barton, N.H. (1997)
Evolution of a species’ range. Am. Nat. 150, 1–23
23 Ronce, O. and Kirkpatrick, M. (2001) When
sources become sinks: migrational meltdown in
heterogeneous habitats. Evolution55, 1520–1531
24 Kawecki, T.J. (2000) Adaptation to marginal
habitats: contrasting influence of the dispersal rate
on the fate of alleles with small and large effects.
Proc. R. Soc. London B Biol. Sci. 267, 1315–1320
25 Proulx, S.R. (1999) Matings systems and the
evolution of niche breadth. Am. Nat. 154, 89–98
26 Kirkpatrick, M. (2000) Reinforcement and
divergence under assortative mating. Proc. R.
Soc. London B Biol. Sci. 267, 1649–1655
27 Laland, K.N. (1994) On the evolutionary
consequences of sexual imprinting. Evolution
48,477–489
28 Wiener, P. and Feldman, M.W. (1993) The effects
of the mating system on the evolution of migration
in a spatially heterogeneous population. Evol.
Ecol.7, 251–269
29 Boughton, D.A. (1999) Empirical evidence for
complex source-sink dynamics with alternative
states in a butterfly metapopulation. Ecology
80,2727–2739
30 Tufto, J. (2001) Effects of releasing maladapted
individuals: a demographic-evolutionary model.
Am. Nat. 158, 331–340
31 Storfer, A. (1999) Gene flow and endangered
species translocations: a topic revisited. Biol.
Conserv.87, 173–180
32 Brown, J. and Kodric-Brown, A. (1977) Turnover
rates in insular biogeography: effect of
immigration and extinction. Ecology58, 445–449
33 Holt, R.D. and Gomulkiewicz, R. (1997) How does
immigration influence local adaptation? A
reexamination of a familiar paradigm. Am. Nat.
149, 563–572
34 Gomulkiewicz, R. et al. (1999) The effects of
density dependence and immigration on local
adaptation and niche evolution in a black-hole
sink environment. Theor. Popul. Biol. 55, 283–296
35 Ingvarsson, P. (2001) Restoration of genetic
variation lost – the genetic rescue hypothesis.
Trends Ecol. Evol.16, 62–63
36 Mallet, J. and Barton, N. (1989) Inference from
clines stabilized by frequency-dependent
selection. Genetics122, 967–976
37 Kruuk, L.E.B. et al. (1999) A comparison of
multilocus clines maintained by environmental
adaptation or by selection against hybrids.
Genetics153, 1959–1971
38 Waser, N.M. (1993) Sex, mating system,
inbreeding, and outbreeding. In The Natural
History of Inbreeding and Outbreeding (Thornhill,
N.W., ed.), pp. 1–16, University of Chicago Press
39 Rundle, H.D. and Whitlock, M.C. (2001) A genetic
interpretation of ecologically dependent isolation.
Evolution55, 198–201
40 Wright, S. (1931) Evolution in mendelian
populations. Genetics16, 97–159
41 Coyne, J.A. et al. (1997) A critique of Sewall
Wright’s shifting balance theory of evolution.
Evolution51, 643–671
42 Ingvarsson, K. (2000) Differential migration from
high fitness demes in the shining fungus beetle,
Phalacrus substriatus. Evolution54, 297–301
43 Barton, N.H. (1986) The effects of linkage and
density-dependant regulation on gene flow.
Heredity57, 415–426
44 Barton, N.H. and Hewitt, G.M. (1985) Analysis of
hybrid zones. Annu. Rev. Ecol. Syst. 16, 113–148
45 Cohan, F.M. et al. (1994) Amelioration of the
deleterious pleiotropic effects of an adaptive
mutation in Bacillus subtilis. Evolution48, 81–95
46 Otto, S.P. and Bourguet, D. (1999) Dominance in
patchy environments. Am. Nat. 153, 561–574
47 Davies, A.G. et al. (1996) Scalloped wings is the
Lucilia cuprima Notch homologue and a candidate
for the Modifierof fitness and asymmetry of
diazinon resistance. Genetics143, 1321–1337
48 Felsenstein, J. (1981) Skepticism towards Santa
Rosalia, or why are there so few kinds of animals.
Evolution35, 124–138
49 Rundle, H.D. and Schluter, D. (1998)
Reinforcement of stickleback mate preferences:
sympatry breeds contempt. Evolution52, 200–208
50 Balkau, B. and Feldman, M.W. (1973) Selection
for migration modification. Genetics74, 171–174
51 Kawecki, T.J. and Stearns, S.C. (1993) The
evolution of life histories in spatially
heterogeneous environments – optimal reaction
norms revisited. Evol. Ecol. 7, 155–174
52 Zhivotovsky, L.A. et al. (1996) On the evolution of
phenotypic plasticity in a spatially heterogeneous
environment. Evolution50, 547–558
53 Kimura, M. (1983) The Neutral Theory of
Molecular Evolution, Cambridge University Press
54 Olivieri, I. et al. (1995) Metapopulation genetics
and the evolution of dispersal. Am. Nat.
146,202–228
55 Venable, D.L. and Brown, J.S. (1988) The
selective interactions of dispersal, dormancy, and
seed size as adaptations for reducing risk in
variable environment. Am. Nat. 131, 360–384
56 Hamilton, W.D. and May, R.M. (1977) Dispersal in
stable habitats. Nature269, 578–581
57 May, R.M. (1979) When to be incestuous. Nature
279, 192–194
58 Gandon, S. and Michalakis, Y. (2001) Multiple causes
of the evolution of dispersal. In Dispersal (Clobert, J.
et al., eds), pp. 155–167, Oxford University Press
59 Montalvo, A.M. et al. (1997) Restoration biology: a
population biology perspective. Rest. Ecol. 5, 277–290
TRENDS in Ecology & Evolution
Vol.17 No.4 April 2002
http://tree.trends.com
189
Review
accelerates the spread of advantageous mutations.
These different forces are likely to interact [58],
and there is no global theoretical synthesis that
includes all of them.
Conclusion
Gene flow tends to oppose the effects of local selection
and thus limits adaptation. However, it can also
replenish the local population and local genetic
variation, which are both pre-requisites for evolution
by natural selection. The relative importance of these
effects is not yet clear. A more thorough evaluation of
the spatial scale of the heterogeneity of selection
among habitats and of the type of selection (directional
or disruptive) is necessary to evaluate the role of gene
flow on adaptation. Finally, manipulating adaptation
by controlling gene flow might have simple and direct
applications, either to disrupt local adaptation of pests
to pesticide [19] or to maintain local adaptation of
agronomical or endangered species [31,59].
... In the case of amphibians, the movement of individuals between spatially disjunct wetland habitats can help maintain genetic diversity, increase local evolutionary potential, and ensure population persistence through rescue effects (Joly, 2019;Semlitsch, 2008). However, movement may also increase extinction risk by facilitating the transmission of virulent diseases, such as chytridiomycosis (Kolby et al., 2015;Muths et al., 2018) and ranavirus (Tornabene et al., 2018), and may erode local adaptation through "gene swamping" (Lenormand, 2002;Tigano & Friesen, 2016). Understanding amphibian dispersal is important for determining the spatial scale at which populations operate and for designing effective disease and habitat management strategies to combat recent species declines (Bailey & Muths, 2019;Converse et al., 2017;Gerber et al., 2018;Smith & Green, 2005). ...
... Our study adds to the growing body of literature suggesting that many wetland-breeding amphibian species are more vagile than previously assumed (e.g., Denoël et al., 2018;Muths et al., 2018;Patrick et al., 2008;Petranka & Holbrook, 2006), particularly within breeding seasons. This movement of individuals can maintain genetic diversity, increase local evolutionary potential (Joly, 2019;Semlitsch, 2008, though see Lenormand, 2002;Tigano & Friesen, 2016), but also facilitate disease transmission (Kolby et al., 2015;Muths et al., 2018;Tornabene et al., 2018). The emergent fungal pathogen Batrachochytrium salamandrivorans (Bsal, described in 2013; Martel et al., 2013) poses a significant conservation risk because of its high pathogenicity to salamander species (Martel et al., 2014). ...
Article
Full-text available
How animals move across space and time in heterogeneous landscapes has important implications for the conservation of imperiled and dispersal‐limited taxa, such as amphibians. Wetland‐breeding amphibians are thought to exhibit strong site fidelity, but there is growing evidence that many species are more vagile than previously assumed. Few studies have quantified breeding dispersal probabilities or distances while also accounting for observational uncertainty, which means that resultant estimates could be biased or confounded with sampling intensity and other demographic rates (e.g., site‐ and time‐specific detection and survival probabilities). Here, we analyze data from a 6‐year capture–mark–recapture study on adult spotted salamanders (Ambystoma maculatum) conducted at 12 wetlands in central Pennsylvania to estimate population dynamics and population structure at multiple scales. We used a multistate, hidden Markov estimator to quantify sex‐specific site fidelity and breeding dispersal as a function of Euclidean distance between wetlands while accounting for imperfect detection. We estimated short‐timescale movements (i.e., those that occur within a single breeding season) and longer timescale movements (i.e., those that occur among breeding seasons) to determine if dispersal rates and distances differed. We found that interannual site fidelity of males varied among wetlands and was positively associated with population density. Females exhibited higher interannual site fidelity and dispersed further than males between breeding seasons. Within breeding seasons, we found that up to 6% of males dispersed to a new wetland each day. Our study is the first to directly compare amphibian breeding dispersal probabilities and distances at multiple scales and provides a robust framework for improving inference on the spatial and temporal patterns of amphibian movement.
... Improve or maintain habitat quality and connectivity; expand suitable ecological space for population(s); assess the need for, and risk of, mixing populations Gene flow Habitat connectivity; geographic distance between populations; movement studies of marked animals Gene flow that is too high can decrease local adaptation; gene flow that is too low can decrease genetic diversity and potentially rates of adaptation. The spatial structure of connections among geographically separate populations affects the extent to which gene flow supports adaptive potential rather than only countering genetic drift, and so on (Lenormand, 2002) Increase connectivity among populations through corridor management; deliberate translocations to increase gene flow among disconnected populations ...
... Exceptions include the potential for enabling disease transmission and epidemic risk (Hess, 1996). Furthermore, increasing connectivity to areas where local adaptations may be important, such as areas mentioned above (climate refugia at the trailing edge of the geographic range, leading edge of the range, environmental gradients), has the potential to facilitate movement of maladapted individuals to those regions (i.e., gene swamping; Lenormand, 2002). A careful approach is to consider population size and genetic distinctiveness. ...
Article
Full-text available
Resource managers have rarely accounted for evolutionary dynamics in the design or implementation of climate change adaptation strategies. We brought the research and management communities together to identify challenges and opportunities for applying evidence from evolutionary science to support on‐the‐ground actions intended to enhance species' evolutionary potential. We amalgamated input from natural‐resource practitioners and interdisciplinary scientists to identify information needs, current knowledge that can fill those needs, and future avenues for research. Three focal areas that can guide engagement include: (1) recognizing when to act, (2) understanding the feasibility of assessing evolutionary potential, and (3) identifying best management practices. Although researchers commonly propose using molecular methods to estimate genetic diversity and gene flow as key indicators of evolutionary potential, we offer guidance on several additional attributes (and their proxies) that may also guide decision‐making, particularly in the absence of genetic data. Finally, we outline existing decision‐making frameworks that can help managers compare alternative strategies for supporting evolutionary potential, with the goal of increasing the effective use of evolutionary information, particularly for species of conservation concern. We caution, however, that arguing over nuance can generate confusion; instead, dedicating increased focus on a decision‐relevant evidence base may better lend itself to climate adaptation actions.
... • la migration : cette force évolutive permet le déplacement de matériel génétique d'une population-source vers une population-puits au-delà des barrières isolant ces populations (McDonald and Linde, 2002;McDonald, 2004). Elle peut conduire à l'introduction de nouveaux allèles dans la population-puits et s'opposer à la sélection naturelle en limitant l'adaptation locale (Lenormand, 2002). Dans le cas particulier des émergences de bactéries pathogènes, la notion de réservoir-source et réservoir-puits est également utilisée pour décrire les migrations. ...
Thesis
Les pathogènes de cultures représentent une menace pour l'Homme depuis les débuts de l'Agriculture. Afinde mieux comprendre les maladies actuelles de cultures et prévenir les épidémies futures, il estindispensable de comprendre les facteurs sous-jacents à l'émergence, l'adaptation et la diffusion desagents pathogènes. De récents développements méthodologiques dans le domaine de l’épidémiologiemoléculaire permettent désormais de reconstruire les dynamiques spatio-temporelles des maladies avecune précision accrue. Alors que la majorité des études précédemment réalisées se sont entièrementappuyées sur l'échantillonnage d'individus contemporains datant des quelques dernières décennies,l’avènement de la paléogénomique permet désormais de reconstruire les génomes historiques de pathogènesdatant de plusieurs siècles et ainsi d'étudier leur histoire évolutive avec une plus grande précision. Afind’évaluer l’apport de la paléogénomique dans l’étude de l’émergence et de l’évolution de pathogènes decultures, nous avons choisi Xanthomonas citri pathovar citri (Xci), la bactérie pathogène responsabledu chancre asiatique des agrumes, comme modèle d’étude. Dans un premier temps nous avons optimiséles protocoles moléculaires et les pipelines bio-informatiques permettant de séquencer et reconstruire aumieux des génomes historiques de Xci à partir de spécimens d’herbiers. Dans ce contexte, nous noussommes particulièrement attaché à étudier les patrons de dégradations de l’ADN dont la mesure estindispensable pour l’authentification des génomes historiques. Dans un second temps, nous avonsprécisé l’histoire de l’émergence de Xci à une échelle locale, celle des îles du sud-ouest de l’océan Indien(SOOI) grâce à l’analyse détaillée du premier génome historique de bactérie pathogène reconstruit à partird’un échantillon d’herbier datant de 1937. Finalement, nous avons significativement amélioré lareconstruction de l’origine et de la diversification de Xci à l’échelle mondiale par l’analyse combinée des13 génomes historiques générés durant cette thèse et d’une collection de génomes modernesreprésentative de la diversité génétique globale du pathogène. Nos travaux soulignent l’importance desdonnées historiques dans la reconstruction de l’histoire évolutive d’agents pathogènes de cultures,valorisant les collections naturalistes et générant des connaissances ayant le potentiel d’optimiser lesstratégies de lutte et de surveillance des épidémies actuelles et de mieux prédire les épidémies futures.
... In such cases, physical barriers limit gene flow, preventing and restricting connectivity and communication among populations, leading to higher rates of genetic drift and inbreeding. Meanwhile, the loss of genetic diversity and inbreeding depression resulted in lower fitness and less evolutionary potential (Lenormand, 2002). ...
Article
Climate oscillation and its synergistic impacts on habitat fragmentation have been identified as threatening the survival of some extant species. However, the mechanisms by which semi-aquatic insects impacted by such events remain poorly understood. Herein, we studied the largest water strider in the world, Gigantometra gigas, to explore the effect of these two factors on its evolutionary history. The sequences of mitogenomic and nrDNA cluster were utilized to reconstruct phylogenetic relationship among G. gigas populations and its demographic history. Mitochondrial genes were separately reconstructed topologies of that populations and detected remarkable differences. We found that G. gigas populations conform to the isolation-by-distance model, and decline occurred at about 120 ka, which was probably influenced by the climate change during the late Pleistocene and eventually maintained a small effective population size (Ne) around 85,717. The populations in Guangdong Province of China are worthy of note in that they exhibit low genetic diversity, a small Ne around 18,899 individuals, and occupy an area with little suitable future habitat for G. gigas. This work recommends that conservation efforts are implemented to ensure the long-term survival of small G. gigas populations, and notes that further evaluation of their extinction risk under the impacts of human activities is required.
... The substantial increase in domestic introgression over the span of only a few decades does not bode well for safeguarding the wild allelic diversity of G. gallus [20,47,48]. This intensification of domestic admixture is likely driven by anthropogenic change to the native environment of G. gallus, which has experienced a high level of habitat loss following the relentless human encroachment into tropical Asia's last wilderness areas in recent decades [49][50][51]. ...
Article
Full-text available
Human activities have precipitated a rise in the levels of introgressive gene flow among animals. The investigation of conspecific populations at different time points may shed light on the magnitude of human-mediated introgression. We used the red junglefowl Gallus gallus , the wild ancestral form of the chicken, as our study system. As wild junglefowl and domestic chickens readily admix, conservationists fear that domestic introgression into junglefowl may compromise their wild genotype. By contrasting the whole genomes of 51 chickens with 63 junglefowl from across their natural range, we found evidence of a loss of the wild genotype across the Anthropocene. When comparing against the genomes of junglefowl from approximately a century ago using rigorous ancient-DNA protocols, we discovered that levels of domestic introgression are not equal among and within modern wild populations, with the percentage of domestic ancestry around 20–50%. We identified a number of domestication markers in which chickens are deeply differentiated from historic junglefowl regardless of breed and/or geographic provenance, with eight genes under selection. The latter are involved in pathways dealing with development, reproduction and vision. The wild genotype is an allelic reservoir that holds most of the genetic diversity of G . gallus , a species which is immensely important to human society. Our study provides fundamental genomic infrastructure to assist in efforts to prevent a further loss of the wild genotype through introgression of domestic alleles.
... Although finding many small-effect loci may be expected in studies of phenotypic evolution (Rockman, 2012), many analyses of adaptation and speciation have found distributions of effect sizes skewed toward larger effects (Griswold, 2006;Lenormand, 2002). ...
Article
Understanding the phenotypic and genetic architecture of reproductive isolation is a longstanding goal of speciation research. In several systems, large‐effect loci contributing to barrier phenotypes have been characterized, but such causal connections are rarely known for more complex genetic architectures. In this study, we combine ‘top‐down’ and ‘bottom‐up’ approaches with demographic modeling toward an integrated understanding of speciation across a monkeyflower hybrid zone. Previous work suggests that pollinator visitation acts as a primary barrier to gene flow between two divergent red‐ and yellow‐flowered ecotypes of Mimulus aurantiacus. Several candidate isolating traits and anonymous SNP loci under divergent selection have been identified, but their genomic positions remain unknown. Here, we report findings from demographic analyses that indicate this hybrid zone formed by secondary contact, but that subsequent gene flow was restricted by widespread barrier loci across the genome. Using a novel, geographic cline‐based genome scan, we demonstrate that candidate barrier loci are broadly distributed across the genome, rather than mapping to one or a few ‘islands of speciation.’ Quantitative trait locus (QTL) mapping reveals that most floral traits are highly polygenic, with little evidence that QTL co‐localize, indicating that most traits are genetically independent. Finally, we find little evidence that QTL and candidate barrier loci overlap, suggesting that some loci contribute to other forms of reproductive isolation. Our findings highlight the challenges of understanding the genetic architecture of reproductive isolation and reveal that barriers to gene flow aside from pollinator isolation may play an important role in this system.
... Animal dispersal is a critical factor in maintaining biodiversity as it is the process underlying gene flow and genetic exchange between populations. Gene flow is an important evolutionary process as it maintains or replenishes the genetic diversity of a population (Lenormand 2002). Restricted gene flow results in population differentiation, with the distribution of genetic variation becoming limited and compartmentalised across the landscape (Moodley et al. 2017). ...
Article
Full-text available
The leopard (Panthera pardus) is facing the threat of continued population decline across its range. In order to inform more effective conservation management programs, genetic information is needed from leopard populations that persist in previously unstudied, isolated and highly fragmented protected areas. The aim of this study was to explore the population structure and genetic diversity of leopard populations across the Mpumalanga province of South Africa. We collected a total of 33 leopard samples from four major locations along a west to east transect across the province. We analysed 17 polymorphic microsatellites and two regions of the mitochondrial genome (NADH-5 and Cytb) to determine the genetic structure of the leopard population in the province. We also calculated genetic diversity indices and explored gene flow in the region. We found that while there is gene flow occurring across the province, the population was genetically structured. We identified two major population units that we describe as ‘West Mpumalanga’ and ‘East Mpumalanga’. Gene flow was moderate between the two populations and we found very high genetic diversity levels compared to other leopard populations previously studied in South Africa. From a conservation perspective, our results show that gene flow is still occurring across seemingly isolated leopard populations that exist in fragmented landscapes, highlighting the importance of all leopard populations in South Africa. Management authorities need to focus conservation efforts on maintaining corridors between regions that are suitable for leopard occupancy and work closely with human settlements to minimise human-leopard conflicts.
Article
Climate change is altering species ranges, and relative abundances within ranges, as populations become differentially adapted and vulnerable to the climates they face. Understanding present species ranges, whether species harbour and exchange adaptive variants, and how variants are distributed across landscapes undergoing rapid change, is therefore crucial to predicting responses to future climates and informing conservation strategies. Such insights are nonetheless lacking for most species of conservation concern. We assess genomic patterns of neutral variation, climate adaptation, and climate vulnerability (offsets in predicted distributions of putatively adaptive variants across present and future landscapes) for sister foundation species, the marine tubeworms Galeolaria caespitosa and Galeolaria gemineoa, in a sentinel region for climate change impacts. We find that species are genetically isolated despite uncovering sympatry in their ranges, show parallel and nonparallel signals of thermal adaptation on spatial scales smaller than gene flow across their ranges, and are predicted to face different risks of maladaptation under future temperatures across their ranges. Our findings have implications for understanding local adaptation in the face of gene flow, and generate spatially explicit predictions for climatic disruption of adaptation and species distributions in coastal ecosystems that could guide experimental validation and conservation planning.
Article
Adaptive divergence occurs even between insufficiently isolated populations when there is a great difference in environments between their habitats. Individuals present in an intermediate zone of the two divergent populations are expected to have an admixed genetic structure due to gene flow. A selective pressure that acts on the genetically admixed individuals may limit the gene flow and maintain the adaptive divergence. Here, we addressed a question whether selection occurs in the genetically admixed individuals between two divergent populations. Arabidopsis halleri is a perennial montane plant, which has clear phenotypic dimorphisms between highland and lowland habitats in Mt. Ibuki, central Japan. We obtained the whole-genome sequences of Arabidopsis halleri plants along an altitudinal gradient of 359–1,317 m with a high spatial resolution (mean altitudinal interval of 20 m). We found a zone where the highland and lowland genes were mixing (intermediate subpopulation). In the intermediate subpopulation, we identified 5 and 13 genome regions, which included 3 and 8 genes, that had a high frequency of alleles that are accumulated in highland and lowland subpopulations, respectively. In addition, we also found that the frequency of highland alleles of these selected genome regions was smaller in the lowland subpopulation compared with that of the non-selected regions. These results suggest that the selection in the intermediate subpopulation might limit the gene flow and contribute to the adaptive divergence between altitudes. We also identified 7 genome regions that had low heterozygote frequencies in the intermediate subpopulation. We conclude that different types of selection in addition to gene flow occur at the intermediate altitude and shape the genetic structure across altitudes.
Article
Full-text available
Epigenetic modifications are thought to be one of the molecular mechanisms involved in plastic adaptive responses to environmental variation. However, studies reporting associations between genome-wide epigenetic changes and habitat-specific variations in life history traits (e.g., lifespan, reproduction) are still scarce, likely due to the recent application of methylome resequencing methods to non-model species. In this study, we examined associations between whole genome DNA methylation and environmentally driven life history variation in two lineages of a marine fish, the capelin (Mallotus villosus), from North America and Europe. In both lineages, capelin harbour two contrasting life history tactics (demersal vs. beach-spawning). Performing whole genome and methylome sequencing, we showed that life history tactics are associated with epigenetic changes in both lineages, though the effect was stronger in European capelin. Genetic differentiation between the capelin harbouring different life history tactics was negligible but we found genome-wide methylation changes in both lineages. We identified 9,125 European and 199 North American differentially methylated regions (DMRs) due to life history. GO enrichment analysis for both lineages revealed an excess of terms related to neural function. Our results suggest that environmental variation causes important epigenetic changes that are associated with contrasting life history tactics in lineages with divergent genetic backgrounds, with variable importance of genetic variation in driving epigenetic variation. Our study emphasizes the potential role of genome-wide epigenetic variation in adaptation to environmental variation.
Article
A model of gene flow and selection in two linked loci is analyzed. The problems considered are the effects of linkage on the clines in frequencies at the two loci and the role of gene flow in producing linkage disequilibrium between the loci. Also, the possible significance of linkage as a mechanism for permitting a population of "track" spatial changes in the environment is considered. The results are that when the recombination fraction between the loci is of the same order of magnitude as the selection coefficients or smaller, then linkage is important in determining the gene frequencies and a substantial amount of linkage disequilibrium is present in the cline. Depending on the spatial pattern of selection on the two loci, linkage can either decrease or increase a population's response to local selection.
Article
Detailed studies of reproductive isolation and how it varies among populations can provide valuable insight into the mechanisms of speciation. Here we investigate how the strength of premating isolation varies between sympatric and allopatric populations of threespine sticklebacks to test a prediction of the hypothesis of reinforcement: that interspecific mate discrimination should be stronger in sympatry than in allopatry. In conducting such tests, it is important to control for ecological character displacement between sympatric species because ecological character divergence may strengthen prezygotic isolation as a by-product. We control for ecological character displacement by comparing mate preferences of females from a sympatric population (benthics) with mate preferences of females from two allopatric populations that most closely resemble the sympatric benthic females in ecology and morphology. No-choice mating trials indicate that sympatric benthic females mate less readily with heterospecific (limnetic) than conspecific (benthic) males, whereas two different populations of allopatric females resembling benthics show no such discrimination. These differences demonstrate reproductive character displacement of benthic female mate choice. Previous studies have established that hybridization between sympatric species occurred in the past in the wild and that hybrid offspring have lower fitness than either parental species, thus providing conditions under which natural selection would favor individuals that do not hybridize. Results are therefore consistent with the hypothesis that female mate preferences have evolved as a response to reduced hybrid fitness (reinforcement), although direct effects of sympatry or a biased extinction process could also produce the pattern. Males of the other sympatric species (limnetics) showed a preference for smaller females, in contrast to the inferred ancestral preference for larger females, suggesting reproductive character displacement of limnetic male mate preferences as well.
Article
We investigate the interplay between gene flow and adaptation in peripheral populations of a widespread species. Models are developed for the evolution of a quantitative trait under clinally varying selection in a species whose density decreases from the center of the range to its periphery. Two major results emerge. First, gene flow from populations at the range center can be a strong force that inhibits peripheral populations from evolving to their local ecological optima. As a result, peripheral populations experience persistent directional selection. Second, response to local selection pressures can cause rapid and substantial evolution when a peripheral population is isolated from gene flow. The amount of evolutionary change depends on gene flow, selection, the ecological gradient, and the trait's heritability. Rapid divergence can also occur between the two halves of a formerly continuous population that is divided by a vicariant event. A general conclusion is that disruption of gene flow can cause evolutionary divergence, perhaps leading to speciation, in the absence of contributions from random genetic drift.
Article
A genetic model for the dynamics of a quantitative trait is analyzed in terms of gene frequencies, linkage disequilibria, and environmental effects on the trait. In a randomly mating population, at each generation progeny move to niches where they are subject to weak Gaussian selection on the trait, with different fitness levels in the different niches. Initially, the variability of the trait is due to additive loci with heterozygous homeostasis. The evolution of plasticity is then described in terms of the invasion of the population by genetic modifiers that may epistatically affect the trait, its optimum in each niche, the strengths of selection, and other parameters characteristic of the niches. We show that the evolution of trait means within niches depends on the overall evolution in the whole system, and in general, optimum phenotypic values are not attained. The reaction norm and genotype-environment interaction may evolve even if the only effects of the modifier are on individual rates of dispersal, or on fitness effects resulting from the different environments in the different niches; this evolution does not require that the modifier affect parameters that influence the values of the trait. It is conjectured that in the least frequently reached niches with low fitness levels, the deviations from the trait optima should be larger than those in more commonly experienced and less stringent niches. Our analysis makes explicit the different contribution of between- and within-niche effects on the evolutionary dynamics of phenotypic plasticity in heterogeneous environments.
Article
The deleterious pleiotropic effects of an adaptive mutation may be ameliorated by one of two modes of evolution: (1) by replacement, in which an adaptive mutation with harmful pleiotropic effects is replaced by one that confers an equal benefit but at less cost; or (2) by compensatory evolution, in which natural selection favors modifiers at other loci that compensate for the deleterious effects of the mutant allele. In this study, we have measured the potential of these two modes of evolution to ameliorate the deleterious pleiotropic effects of resistance to the antibiotic rifampicin in the soil bacterium Bacillus subtilis. One approach was to measure the fitness cost of a series of spontaneous rifampicin-resistance mutations from each of several strains. The potential for amelioration by the replacement mode was estimated by the variation in fitness cost among the mutants of a single strain. Another approach was to introduce a series of different rifampicin-resistance alleles into a diversity of strains, and to measure the fitness cost of rifampicin resistance for each allele-by-strain combination. The potential for amelioration by the replacement mode was estimated by the variation in fitness costs among rifampicin-resistance alleles; the potential for compensatory evolution was estimated by variation in the fitness cost of rifampicin resistance among strains. This study has shown that the cost of rifampicin resistance may be ameliorated by both the compensatory and replacement modes.
Article
We evaluate Sewall Wright's three-phase "shifting balance" theory of evolution, examining both the theoretical issues and the relevant data from nature and the laboratory. We conclude that while phases I and II of Wright's theory (the movement of populations from one "adaptive peak" to another via drift and selection) can occur under some conditions, genetic drift is often unnecessary for movement between peaks. Phase III of the shifting balance, in which adaptations spread from particular populations to the entire species, faces two major theoretical obstacles: (1) unlike adaptations favored by simple directional selection, adaptations whose fixation requires some genetic drift are often prevented from spreading by barriers to gene flow; and (2) it is difficult to assemble complex adaptations whose constituent parts arise via peak shifts in different demes. Our review of the data from nature shows that although there is some evidence for individual phases of the shifting balance process, there are few empirical observations explained better by Wright's three-phase mechanism than by simple mass selection. Similarly, artificial selection experiments fail to show that selection in subdivided populations produces greater response than does mass selection in large populations. The complexity of the shifting balance process and the difficulty of establishing that adaptive valleys have been crossed by genetic drift make it impossible to test Wright's claim that adaptations commonly originate by this process. In view of these problems, it seems unreasonable to consider the shifting balance process as an important explanation for the evolution of adaptations.
Article
Predation is an important selective force that influences animal color patterns. Some larval populations of the streamside salamander, Ambystoma barbouri, inhabit streams with fish predators. Other larval salamanders are found in shallow, ephemeral streams that are predator-free. Quantitative melanophore cell counts and estimates of percent body area pigmented indicated that larval coloration is strongly correlated with stream type. Larvae that coexist with fish tend to be lighter than larvae from streams that are Ashless and ephemeral. Two approaches demonstrated that lightly pigmented salamander larvae better match the common background in relatively permanent streams and are less conspicuous to fish than dark larvae. First, using a model based on the spectral sensitivity of the fish and reflectance properties of salamanders and natural stream backgrounds, we showed that light larvae are three times more cryptic than dark larvae on rocks. Second, lighter larvae had higher survival than darker salamanders on rocks in a predator- choice experiment. It is not clear why larvae in ephemeral streams are darker. Larvae in ephemeral streams should be active to feed and develop rapidly and reach sufficient size to metamorphose before seasonal drying. Several hypotheses may explain why larvae tend to be darker in ephemeral streams, such as increased thermoregulatory ability, better screening of ultraviolet radiation (in these shallower streams), or better background matching to terrestrial predators. Among populations where salamander larvae coexist with fish, there are differences in relative crypsis. Larvae from populations with fish and relatively high gene flow from ephemeral populations (where larvae are dark) tend to be darker (with more melanophores) and more conspicuous to predators than those from more genetically isolated populations, where larvae are lighter and more cryptic. These differences illustrate the role of gene flow as a constraint to adaptive evolution.
Article
Predation is an important selective force that influences animal color patterns. Some larval populations of the streamside salamander, Ambystoma barbouri, inhabit streams with fish predators. Other larval salamanders are found in shallow, ephemeral streams that are predator-free. Quantitative melanophore cell counts and estimates of percent body area pigmented indicated that larval coloration is strongly correlated with stream type. Larvae that coexist with fish tend to be lighter than larvae from streams that are fishless and ephemeral. Two approaches demonstrated that lightly pigmented salamander larvae better match the common background in relatively permanent streams and are less conspicuous to fish than dark larvae. First, using a model based on the spectral sensitivity of the fish and reflectance properties of salamanders and natural stream backgrounds, we showed that light larvae are three times more cryptic than dark larvae on rocks. Second, lighter larvae had higher survival than darker salamanders on rocks in a predator- choice experiment. It is not clear why larvae in ephemeral streams are darker. Larvae in ephemeral streams should be active to feed and develop rapidly and reach sufficient size to metamorphose before seasonal drying. Several hypotheses may explain why larvae tend to be darker in ephemeral streams, such as increased thermoregulatory ability, better screening of ultraviolet radiation (in these shallower streams), or better background matching to terrestrial predators. Among populations where salamander larvae coexist with fish, there are differences in relative crypsis. Larvae from populations with fish and relatively high gene flow from ephemeral populations (where larvae are dark) tend to be darker (with more melanophores) and more conspicuous to predators than those from more genetically isolated populations, where larvae are lighter and more cryptic. These differences illustrate the role of gene flow as a constraint to adaptive evolution.
Article
We evaluate Sewall Wright's three-phase 'shifting balance' theory of evolution, examining both the theoretical issues and the relevant data from nature and the laboratory. We conclude that while phases I and II of Wright's theory (the movement of populations from one 'adaptive peak' to another via drift and selection) can occur under some conditions, genetic drift is often unnecessary for movement between peaks. Phase III of the shifting balance, in which adaptations spread from particular populations to the entire species, faces two major theoretical obstacles: (1) unlike adaptations favored by simple directional selection, adaptations whose fixation requires some genetic drift are often prevented from spreading by barriers to gene flow; and (2) it is difficult to assemble complex adaptations whose constituent parts arise via peak shifts in different demes. Our review of the data from nature shows that although there is some evidence for individual phases of the shifting balance process, there are few empirical observations explained better by Wright's three-phase mechanism than by simple mass selection. Similarly, artificial selection experiments fail to show that selection in subdivided populations produces greater response than does mass selection in large populations. The complexity of the shifting balance process and the difficulty of establishing that adaptive valleys have been crossed by genetic drift make it impossible to test Wright's claim that adaptations commonly originate by this process. In view of these problems, it seems unreasonable to consider the shifting balance process as an important explanation for the evolution of adaptations.