Nonneutral Admixture of Immigrant Genotypes in African Drosophila
melanogaster Populations from Zimbabwe
Maximilian Kauer, Daniel Dieringer, and Christian Schlötterer
Institut für Tierzucht und Genetik, Vienna, Austria
Drosophila melanogaster originated in Africa and colonized the rest of the world only recently (approximately 10,000 to
15,000 years ago). Using 151 microsatellite loci, we investigated patterns of gene flow between African D. melanogaster
populations representing presumptive ancestral variation and recently colonized European populations. Although we
detected almost no evidence for alleles of non-African ancestry in a rural D. melanogaster population from Zimbabwe,
an urban population from Zimbabwe showed evidence for admixture. Interestingly, the degree of admixture differed
among chromosomes. X chromosomes of both rural and urban populations showed almost no non-African ancestry, but
the third chromosome in the urban population showed up to 70% of non-African alleles. When chromosomes were
broken into contingent microsatellite blocks, even higher estimates of admixture and significant heterogeneity in
admixture was observed among these blocks. The discrepancy between the X chromosome and the third chromosome is
not consistent with a neutral admixture hypothesis. The higher number of European alleles on the third chromosome
could be due to stronger selection against foreign alleles on the X chromosome or to more introgression of (beneficial)
alleles on the third chromosome.
Introduction
Drosophila melanogaster is assumed to have an
African origin and colonized the rest of the world only
recently (David and Capy 1988). Morphological, genetic
and biogeographic data suggested that the European
continent was colonized after the last glaciation 10,000
to 15,000 years ago. North America was mainly colonized
from Europe only about 100 years ago (David and Capy
1988). Recent microsatellite data suggest that some North
American D. melanogaster populations experienced admixture of African alleles (Caracristi and Schlötterer
2003), which seems to imply an additional colonization
route from Africa (David and Capy 1988). Sub-Saharan
African populations harbor the highest levels of genetic
variability, whereas polymorphism outside of Africa is
considerably lower (Andolfatto 2001; Kauer et al. 2002;
Schlötterer and Harr 2002; Caracristi and Schlötterer
2003). Although African and non-African populations are
substantially divergent in phenotype and genotype (Begun
and Aquadro 1993), the molecular variation detected
in non-African populations is mainly a subset of the
variability present in Africa (Andolfatto 2001; Schlötterer
and Harr 2002). Within (sub-Saharan) Africa and among
non-African populations only, low levels of differentiation
were detected (Begun and Aquadro 1995; Caracristi and
Schlötterer 2003). Although the maintenance of the differentiation between African and non-African flies could
be explained by geographic barriers, such as the Sahara
Desert or the sea, with the increase in transportation of
goods and human travel, a commensal species, such as
D. melanogaster, could easily cross these barriers (Capy et
al. 2000). D. melanogaster populations from Africa and
other continents differ in various traits such as morphology, ethanol tolerance, or cuticular hydrocarbons (Vouidibio et al. 1989; Capy, Pla, and David 1993; Takahashi
Key words: Drosophila melanogaster, gene flow, selection, habitat
adaptation, antagonistic pleiotropy.
E-mail: [email protected].
Mol. Biol. Evol. 20(8):1329–1337. 2003
DOI: 10.1093/molbev/msg148
Molecular Biology and Evolution, Vol. 20, No. 8,
Ó Society for Molecular Biology and Evolution 2003; all rights reserved.
et al. 2001). Therefore, it can be assumed that African and
non-African flies are sufficiently divergent that some
biological reproductive barrier, such as mating preferences
or hybrid disadvantage, has already emerged. Divergence
could be due to neutral drift or adaptation to different
habitats (Cooper and Lenski 2000; Hawthorne and Via
2001; Hendry 2001). Although a mating barrier has been
described (Wu et al. 1995; Alipaz, Wu, and Karr 2001; see
also Discussion) flies from Zimbabwe and non-African
flies can be easily crossed in the laboratory, and it is not
clear to what extent this behavior might impede crossings
in nature. To get a more detailed picture of population
differentiation between African and non-African flies, we
inferred naturally occurring gene flow between Zimbabwean and European populations of D. melanogaster using
presumably neutral microsatellite variation. Significant
differences were observed between the third and the X
chromosome. While evidence for admixture of European
alleles was detected for the Zimbabwe third chromosomes
in an urban population, the X chromosome in Zimbabwe
showed almost no evidence for admixture.
Material and Methods
Microsatellites
We typed 77 and 74 microsatellites on the third
chromosome and X chromosome, respectively. Microsatellites can be assumed to be very rarely affected by
selection and can therefore be regarded as neutral markers.
Genotyping of microsatellite loci followed standard protocols (Schlötterer and Zangerl 1999). Primer sequences,
annealing temperatures, repeat motifs, and cytological
position of all loci are available as online Supplementary
Material at the journal’s Web site.
Fly Strains
Zimbabwe flies were sampled from two locations,
Sengwa Wildlife Reserve (ZS) or the capital city of
Zimbabwe, Harare (ZH), and were kindly provided by C.
F. Aquadro and C.-I. Wu. The lines established from both
populations were propagated in the same laboratories,
1329
1330 Kauer et al.
Table 1
Genetic Differentiation Between Populations
X Chromosome
Third Chromosome
FST/D
Rome
Friedrichshafen
ZH
ZS
FST/D
Naples
Katovice
ZH
ZS
Rome
Friedrichshafen
ZH
ZS
—
0.209
0.649
0.674
0.0396**
—
0.648
0.669
0.258**
0.249**
—
0.406
0.272**
0.265**
0.003 ns
—
Naples
Katovice
ZH
ZS
—
0.173
0.383
0.473
0.0368**
—
0.402
0.493
0.141**
0.125**
—
0.289
0.2158**
0.213**
0.0256*
—
NoTE.—Country of origin of populations; Germany: Friedrichshafen; Italy: Rome and Naples; Poland: Katovice; Zimbabwe: Harare (ZH) and Sengwa (ZS). Upper
triangles of tables are FST values; lower triangles are genetic distances (D) (1 proportion of shared alleles). **P , 0.01; ns indicates nonsignificant.
making them equally exposed to possible contamination
by other D. melanogaster lines. To account for inbreeding
effects in the Zimbabwean isofemale lines, we randomly
discarded one allele from heterozygous individuals.
European flies were from Italy (Naples and Rome,
collected 1998 and 2001), Poland (Katovice, collected by
Jacek Gorczyca, 2000) and Germany (Friedrichshafen,
collected by B. Harr, 1998). Naples and Katovice flies
were typed for the third chromosome; Rome and
Friedrichshafen flies were typed for the X chromosome.
For each European population, 30 F1 individuals (i.e., first
generation progeny from a single, wild-caught, inseminated female) were used. Note that the fact that different
populations are used for different chromosomes had purely
technical reasons and does not introduce a bias as genetic
differentiation between European populations is very low
among the analyzed populations (table 1).
Measures of Genetic Differentiation
All calculations were done with Microsatellite-Analyzer (MSA) (Dieringer and Schlötterer 2003). Genetic
distances were calculated as 1-proportion of shared alleles.
h-values were calculated as unbiased estimators of FST
(Weir and Cockerham 1984). Significance levels for FSTvalues were calculated by permuting (1,000 times) genotypes among populations. We used the Bonferroni
correction to account for multiple testing (Sokal and
Rohlf 1995).
Interpretation of X Chromosomal and Third
Chromosomal Differences
In this study, we performed a comparison of X
chromosomes and autosomes. This may be complicated
because these chromosomes differ for features like levels
of variability (Andolfatto 2001; Kauer et al. 2002) and the
presence of inversions (Lemeunier and Aulard 1992).
Inversions can reduce the recombination rate on the autosomes, which in turn affects levels of natural variability (Begun and Aquadro 1992). In our study, however,
we analyzed two African populations, which were similar in their X chromosomal (Harare: H ¼ 0.77; Sengwa:
H ¼ 0.76) and autosomal (Harare: H ¼ 0.66; Sengwa:
H ¼ 0.61) levels of variability and also in their allele
distribution. Hence, all the above mentioned differences are
expected to apply to both African populations to the same
extent. Patterns differing between these populations are
therefore unlikely to be caused by these features.
Bayesian Admixture Analysis
The program Structure (Pritchard, Stephens, and
Donnelly 2000) was used to assign individuals to
homogenous clusters (populations) without consideration
of the sampling localities. This program uses a Bayesian
model–based clustering method for multilocus genotypes
to simultaneously determine the most probable number of
homogenous populations in a given data set and assign
individuals to one or more of them. The number of clusters
is inferred by calculating the probability P(X j K) of the
data given a certain prior value of K (number of clusters)
over a number of Monte Carlo Markov Chain (MCMC)
iterations. The posterior probabilities P(K j X) can be
calculated following Bayes’ rule. The clusters are
characterized by different allele frequencies, and, according to their allele distributions, individuals are probabilistically assigned to one or more clusters. The scores of
individuals in the clusters correspond to the probability of
ancestry in any one of them. In this study we assumed
prior values of K from 1 to 5. All calculations shown in
this report are based on 1,000,000 iterations of the
MCMC, after a ‘‘burn-in’’ period of 50,000 iterations.
(The burn-in period are the first iterations of the MCMC,
which are dependent on the start configuration. These
iterations are not incorporated in the final calculation of the
posterior probability [Pritchard, Stephens, and Donnelly
2000].) We ran the program without incorporation of prior
population information. Before extracting definite values,
we ran the program for each prior of K for different
numbers of iterations to check for homogeneity over runs.
Long runs were made to get accurate estimations of
P(X j K) (Pritchard, Stephens, and Donnelly 2000). All
simulations were run under a model that allows for
admixture between populations (Admixture model). We
also used a new, unpublished version of Structure (version
II, available from http://pritch.bsd.uchicago.edu/) to incorporate linkage information between the markers and to
use a different model that accounts for correlation of allele
frequencies among populations due to shared history (D.
Falush et al., personal communication; see also description
of the program available at the above internet site). Genetic
map distances for D. melanogaster were downloaded from
http://flybase.bio.indiana.edu/maps/lk/cytotable.txt. If not
stated otherwise, we used the old version of Structure.
To check for heterogeneity of admixture along the
chromosome, we divided our data for both chromosomes
into five blocks of 14 to 18 adjacent microsatellites. On the
third chromosome, we discarded two loci for this analysis
that were not close to any other microsatellites. The absolute
Nonneutral Admixture in Drosophila melanogaster 1331
Table 2
FST Values and Genetic Distance Between European and
Zimbabwean Populations
Genetic Distanceb
FSTa
All loci
X chromosome loci
Third chromosome loci
a
b
ZS
ZH
ZS
ZH
0.24
0.27
0.22
0.20
0.26
0.14
0.57
0.67
0.47
0.53
0.65
0.40
All FST values are highly significant (P , 0.001).
Genetic distances are 1 proportion of shared alleles.
nucleotide positions (Mb) in the Drosophila sequence
(release 2, available from: http://flybase.bio.indiana.edu/)
and cytological positions (polythene bands) that were
covered by the locus sets X-1 to X-5 and 3-1 to 3-5 on
the respective chromosomes were X chromosome, X-1: 15
loci, Mb 1.31 to 3.52, band 1E-3F; X-2: 15 loci, Mb 3.64 to
5.67, band 3F-5C; X-3: 15 loci, Mb 5.71 to 9.32, band 5C8E; X-4: 15 loci, Mb 9.32 to 14.55, band 8E-12F; X-5: 14
loci, Mb 15.14 to 19.94, band 13C-19C; chromosome 3, 3-1:
18 loci, Mb 3L2.29 to 10.13, band 62C-67D; 3-2: 14 loci,
Mb 3R0.78 to 3.25, band 82C-84D; 3-3: 14 loci, Mb 3R3.64
to 5.87, band 84d-85f; 3-4: 14 loci, Mb 3R5.88 to 16.18,
band 85f-92e; 3-5: 15 loci, Mb 3R17.29 to 26.33, band 93e100a. For each of these blocks, we ran Structure with K ¼ 2
(all other settings as above).
Results
We analyzed 74 and 77 microsatellites mapping to
the X chromosome and third chromosome, respectively.
Consistent with previous studies, we found European and
Zimbabwean flies to be well differentiated (Begun and
Aquadro 1993; Begun and Aquadro 1995; Andolfatto
2001; Caracristi and Schlötterer 2003). The mean FST
between European and Zimbabwean populations is 0.23 (P
, 0.001), and the genetic distance based on the proportion
of shared alleles is 0.57. Population differentiation among
European populations is low but significant (table 1).
Given that the level of differentiation between European
populations is very low compared with the differentiation
between African and non-African populations, this should
not bias the analysis. Within Africa, population differentiation was lower than among European populations (table
1). A closer inspection of the genetic differentiation
between the European populations and two populations
from Zimbabwe indicated that the urban population collected in Harare was less differentiated from the European
populations than the rural population collected at the
Sengwa Wildlife Reserve (table 2).
Since recent studies indicated that the partitioning of
variability differs dramatically among X chromosome and
autosomes in D. melanogaster (Andolfatto 2001; Kauer
et al. 2002), we analyzed both chromosomes separately. As
the differentiation between the European populations is
very low, the analysis of different European populations
for the X chromosome and the third chromosome is not
expected to influence our results. Surprisingly, the genetic
differentiation was almost identical for X-linked micro-
Table 3
Inferring the Number of Populations Using Structure
Third Chromosome
X Chromosome
Ka
ln P(X j K)
P(K j X)b
ln P(X j K)
P(K j X)b
1
2
3
4
5
13321.3
12389.8
12394.3
12396.9
12398.8
0.000
0.988
0.011
0.001
0.000
14180.4
12528.6
12587.9
12581.3
12577.6
0.00
1.00
0.00
0.00
0.00
a
b
Numbers of clusters that were assumed in a run of Structure.
Assuming a uniform prior for K (K 2 [1,2,3,4,5]).
satellites irrespective of whether Sengwa or Harare was
compared with the European populations (tables 1 and 2).
For the third chromosome, however, we found the Harare
population to be more similar to Europeans than the
Sengwa population, irrespective of whether we used FST
or the proportion of shared alleles. Thus, the third chromosome harbors more non-African alleles than the X
chromosome, and this effect is more pronounced in the
Harare population.
We used a Bayesian model–based clustering method
(Pritchard, Stephens, and Donnelly 2000) to investigate the
hypothesis of a different amount of European (‘‘cosmopolitan’’) alleles in the two populations from Zimbabwe
and among X chromosomes and third chromosomes. For
both chromosomes, we obtained the highest posterior
probability for two clusters (K ¼ 2, table 3), which
corresponded to African and European populations. Using
the admixture model of Structure, the clustering method
also estimates the probability of ancestry for each
individual in one of the two groups (African-European).
Whereas all European individuals had very high probabilities of ancestry in only one group (‘‘the European
cluster’’), African individuals were found to score in both
groups (table 4). The probabilities of ancestry of African
individuals in the European cluster were ranging from
0% up to 70% (data not shown, but see Appendix A).
Consistent with FST analysis, we found the probability of
ancestry in the European cluster to differ sharply between
the two Zimbabwean populations as well as between
chromosomes (table 4). The highest scores for Zimbabwean individuals in the European cluster were found on
the third chromosome in the capital Harare, whereas the
rural population, Sengwa, shows significantly less influence of the cosmopolitan type (table 4) (P , 0.01,
Mann-Whitney U test). For the X chromosome, on the
other hand, scores in the European cluster are significantly
lower than for the autosome (table 4) (P , 0.001, MannWhitney U test). Despite not being statistically significant,
a trend for more European alleles in the Harare population
can also be seen on the X chromosome (table 4). As this
analysis was based on a substantially larger number of
non-African chromosomes than African ones (120 versus
35 on chromosome 3 and 120 versus 26 on the X chromosome), we repeated the analysis with a data set in which
a number of European chromosomes were randomly
discarded so that equal numbers of chromosomes in
African and European flies were compared. This reduced
data set resulted in the same difference between the X
1332 Kauer et al.
Table 4
Assignment and Inferred Ancestry of Individuals by
Structure
Third Chromosome
X Chromosome
Strains
Cluster1
Cluster2
Cluster1
Cluster2
Europe
Harare
Sengwa
0.994
0.202
0.055
0.006
0.798
0.945
0.999
0.019
0.005
0.001
0.980
0.994
NoTE.—Summarized output of the Bayesian clustering: values are the mean
probabilities of ancestry in the two clusters for each population and chromosome.
chromosomes and the autosome for the Sengwa and
Harare population (Appendix A). Hence, the Bayesian
model–based clustering method provided strong support
for the presence of a larger number of European alleles on
the third chromosome in the Harare population than
expected by either X chromosomal data or the rural
Sengwa population. The same qualitative differences between chromosomes and populations were observed when
we used a newer version of Structure, which accounts for
linkage among microsatellite loci or the shared history of
African and non-African D. melanogaster populations
(data not shown).
Given the evidence for admixture in the Zimbabwean
population, we were further interested in whether this
pattern is seen across the entire chromosome or if it is
limited to some chromosomal regions. First we analyzed
locus-wise FST values between Harare and Sengwa populations and found a homogenous pattern of low differentiation across loci, and also no locus had a significant
FST value after Bonferroni correction (results not shown).
This result therefore provides no evidence for locusspecific differences between the populations (e.g., due
to different inversion frequencies). We also plotted locuswise FST values between the African and non-African populations according to their chromosomal position. This
analysis showed stochastic fluctuations among loci on both
chromosomes rather than adjacent parts of the chromosomes having more similar FST values than more
distant ones (results not shown). Therefore, this analysis
provided no evidence for spatial clustering of admixture
along the chromosome. A different approach to detect
spatial heterogeneity along the chromosomes is provided by the Bayesian model–based clustering method
in which signals of individuals can be detected separately.
Therefore we ordered the loci on both chromosomes
according to their chromosomal position and divided the
data into five sets, each consisting of 14 to18 adjacent loci
(see Material and Methods). With these data sets we ran
the program Structure assuming K ¼ 2. In the Sengwa
population, the probabilities of Zimbabwean ancestry were
very similar among all sets of loci, irrespective whether
they were located on the X chromosome or on the third
chromosome (fig. 1 and Appendix A). For the Harare
population, differences between the locus sets were more
pronounced, showing the largest variance on the third
chromosome (fig. 1b and Appendix B). The variance of the
individual probabilities of ancestry between the locus sets
is higher than the variance among individuals within sets.
On the level of individuals, the difference in probability of
FIG. 1.—Boxplots of assignment of individuals to the ‘‘European’’
cluster (inferred European ancestry) using different locus sets on the X
chromosome and autosome (actual values can be found in Appendix A).
Boxes include the two interquartiles containing 50% of the data.
Horizontal lines within the boxes represent the median. Vertical lines
lead to highest and lowest values but not including extreme values.
Extreme values were defined as lying outside of 1.5 box lengths and were
represented by black dots. In each figure the five locus sets (L1 to L5) for
the Harare and Sengwa sample are shown. (A) X chromosome. (B) Third
chromosome.
non-African origin is very pronounced. Individual ZH1,
for example, shows 35% and 99% probability of ancestry
in the European cluster for locus sets 3-1 and 3-2, but only
4%, 5%, and 5% for locus sets 3-3, 3-4, and 3-5. Basically
in all locus sets, at least one individual had admixture
values higher than 0.5. Again this result was qualitatively
the same when the linkage model or the correlated allele
frequency model of Structure II was used (data not
shown).
Discussion
Our microsatellite survey provided evidence for the
presence of non-African alleles on the third chromosome in
an African D. melanogaster population collected in Harare,
the capital of Zimbabwe. In contrast, the X chromosome
Nonneutral Admixture in Drosophila melanogaster 1333
harbored almost no non-African alleles. Interestingly, this
discrepancy between X chromosome and third chromosome was not found when a rural population (Sengwa)
was analyzed. Given that the differences between the X
chromosome and the autosome were not detected in both
populations from Zimbabwe, locus-specific and also
chromosome-specific effects causing an erroneous signal
of admixture in Zimbabwe seem unlikely, as these would
result in a consistent pattern in both populations. For
example, a higher power to detect admixture on the third
chromosome than on the X chromosome (e.g., due to more
variability on the third chromosome in non-African populations) should apply to both sample locations from
Zimbabwe.
An issue of special interest are chromosomal inversions known to segregate on the autosomes of D.
melanogaster at high frequencies (Lemeunier and Aulard
1992). The heterogeneity in admixture among individuals
and chromosomal regions (fig. 1b and Appendix B) could, in
principle, be also attributed to chromosomal inversions. For
example, old inversions causing tight linkage among loci on
the third chromosome that are segregating in Zimbabwe
populations but are fixed in European populations (e.g., due
to selection) could be misinterpreted as admixture of nonAfrican alleles. Such a scenario, however, is unlikely to
account for the inferred non-African ancestry in Zimbabwe,
as Sengwa and Harare showed very low differentiation over
all loci (table 1) and also locus-wise FST values did not
provide evidence for differentiated genomic regions. Thus,
if the inversions have been segregating in Zimbabwe for
a long time they should be shared between both populations
and occur at similar frequencies.
As we observed a high fraction of non-African alleles
on the third chromosome only in Harare, this scenario
seems to be not compatible with our data. Nevertheless,
direct proof that Sengwa and Harare populations are not
differentiated with respect to inversions could be only
gained if these populations were karyotyped.
Alternatively, our observation could be explained by
a combination of selection and demography. Whereas in
the rural Sengwa population, Zimbabwean ancestry was
observed for both chromosomes, non-African ancestry up
to 70% was detected on the third chromosome in the
Harare population. In several individuals collected in
Harare, we detected chromosomal segments of almost
pure non-African descent (.80% [Appendix A]). This result suggests that the rural population experiences less
admixture of non-African genotypes than the population
collected at the capital Harare. This is consistent with
more transportation from Europe to the capital than to
the wildlife reserve and also with other studies from D.
melanogaster in Africa that found more ‘‘cosmopolitanlike’’ flies in towns than in the countryside (Vouidibio et
al. 1989; Capy, Pla, and David 1993). Importantly, we also
found a striking difference between the third chromosome
and X chromosome in Harare. As neutral admixture affects
both chromosomes to the same extent, admixture alone is
not sufficient to explain our data.
In the following we discuss different nonneutral
scenarios, which could potentially explain the observed
pattern of admixture. First we discuss the possibility that
European alleles could have a selective advantage in the
Zimbabwean populations (i.e., positive selection for introgressed alleles in Zimbabwe). In contrast, the next
two sections assume that European alleles have a selective
disadvantage in Zimbabwe (i.e., purifying selection against
these alleles in Zimbabwe).
Beneficial Effects of European Alleles
One nonneutral explanation for the different degree of
admixture could be that some of the European alleles could
have a selective advantage in Zimbabwe. A potential example could be insecticide resistance, which has recently
been suggested to have spread throughout worldwide
populations (Daborn et al. 2001). With the fixation of
such beneficial alleles during a selective sweep, a larger
genomic region around such alleles could be fixed in the
African population via hitchhiking (Maynard Smith and
Haigh 1974). Hence, if several beneficial alleles are
located on the third chromosome, this would result in
more admixture on the third chromosome than on the X
chromosome. However, it is not obvious that multiple
beneficial alleles should be located on the third chromosome, but none on the X chromosome. Whether the
different environments from which the African populations were sampled (urban versus rural) affect the degree
of admixture cannot be inferred from our data, as our
sample sizes are rather restricted. The absence of the
European alleles in the Sengwa sample might either be due
to restricted gene flow between Sengwa and Harare or to
selection against European alleles in the rural populations.
Studying larger sample sizes and sample gradients from
the town to the countryside could shed more light on this.
Sexual Selection
Sexual selection assumes that fast evolving traits for
mating preferences could have evolved in different directions in the Zimbabwean and non-African populations.
A secondary contact of populations divergent for mating
preferences could therefore result in assortative mating
and in postmating isolation. Interestingly, such a pattern of
mating preference has been described for the Zimbabwean
populations, the Z-M mating behavior (Wu et al. 1995;
Hollocher et al. 1997a). In brief D. melanogaster
from Zimbabwe (Z) and also from some other African
populations prefer to mate with flies from their own
populations when given the choice between these and so
called ‘‘cosmopolitan flies’’ (M) (i.e., flies from outside of
Africa). This mating difference was found to be asymmetrical in two ways. First only African flies discriminate but
not non-African flies. Second only female African flies
discriminate; males readily mate with cosmopolitan
females. Whereas a wide range of Z-values (degree of mate
discrimination of females) was observed in the Harare
population, the rural Sengwa population had on average
a stronger Z-like (i.e., discriminating) mating behavior
(Hollocher et al. 1997a). Mapping studies indicated that the
assortative mating behavior mapped primarily to the
autosomes, but only small effects were found on the X
chromosome (Hollocher et al. 1997b; Ting, Takahashi, and
1334 Kauer et al.
Appendix A
Assignment of Individuals to the ‘‘European’’ Cluster Using Different Locus Sets on the X Chromosome
Locus Set
Individuals
Mean European
Zh1
Zh2
Zh3
Zh12
Zh13
Zh16
Zh18
Zh19
Zh20
Zh23
Zh26
Zh28
Zh40
Mean Zh
Zs2
Zs6
Zs7
Zs10
Zs11
Zs15
Zs22
Zs24
Zs28
Zs29
Zs30
Zs35
Zs37
Mean Zs
a
X-1
X-2
X-3
X-4
X-5
Mean over Locus Sets
Complete Data Set
Reduced Data Seta
0.997
0.020
0.024
0.027
0.053
0.004
0.010
0.031
0.004
0.006
0.003
0.006
0.011
0.005
0.016
0.019
0.006
0.027
0.013
0.484
0.013
0.015
0.008
0.006
0.014
0.006
0.024
0.015
0.050
0.996
0.115
0.006
0.066
0.035
0.022
0.049
0.228
0.008
0.015
0.014
0.137
0.010
0.759
0.113
0.005
0.012
0.020
0.003
0.005
0.023
0.022
0.006
0.032
0.006
0.009
0.011
0.008
0.012
0.997
0.005
0.035
0.008
0.007
0.132
0.200
0.029
0.005
0.087
0.020
0.012
0.005
0.019
0.044
0.018
0.048
0.023
0.031
0.031
0.006
0.019
0.004
0.005
0.016
0.006
0.264
0.007
0.037
0.997
0.008
0.005
0.005
0.004
0.031
0.003
0.003
0.007
0.004
0.010
0.035
0.003
0.004
0.009
0.016
0.004
0.003
0.005
0.006
0.007
0.006
0.005
0.028
0.004
0.054
0.005
0.003
0.011
0.996
0.010
0.008
0.010
0.009
0.545
0.005
0.005
0.055
0.012
0.026
0.378
0.044
0.011
0.086
0.034
0.006
0.039
0.005
0.006
0.026
0.016
0.005
0.004
0.017
0.012
0.011
0.026
0.016
0.997
0.032
0.016
0.023
0.022
0.147
0.053
0.059
0.016
0.025
0.015
0.114
0.014
0.159
0.053
0.018
0.015
0.022
0.011
0.106
0.015
0.016
0.005
0.015
0.011
0.018
0.063
0.012
0.025
0.999
0.004
0.002
0.037
0.002
0.089
0.005
0.011
0.003
0.003
0.003
0.078
0.003
0.005
0.019
0.006
0.002
0.006
0.001
0.025
0.004
0.006
0.001
0.004
0.002
0.004
0.008
0.002
0.005
0.999
0.005
0.005
0.077
0.007
0.109
0.009
0.011
0.006
0.006
0.007
0.134
0.004
0.006
0.030
0.010
0.004
0.011
0.002
0.033
0.008
0.013
0.002
0.007
0.004
0.010
0.012
0.002
0.009
Data set with the same number of European and Zimbabwean chromosomes.
Wu 2001). More recently evidence for a postmating barrier
was also found as crosses of Z-females with cosmopolitan
males yielded fewer offspring than the reciprocal cross
(Alipaz, Wu, and Karr 2001).
The asymmetry of the M-Z trait makes it difficult to
predict to what extent the mixing of Zimbabwean and
cosmopolitan populations is impaired by this behavior. For
example, the preference of mating in one direction (M
females with Z males) could impose a form of differential
selection (e.g., stronger selection on the introgressed X
chromosome in males [see below]) on the introgression of
X chromosomes versus autosomes. On the other hand, less
admixture could be expected on the third chromosome
because most variance for the mating trait has been
mapped to the autosomes (Hollocher et al. 1997b; Ting,
Takahashi, and Wu 2001). To confirm that the higher
admixture is a general phenomenon of autosomes, rather
than of the third chromosome, we reanalyzed a new data
set consisting of microsatellite variation on the X
chromosome and the second chromosome (Caracristi and
Schlötterer 2003) in Zimbabwe/Harare and non-African
populations. Despite that a smaller number of loci was
characterized, we also found less admixture in Harare
on the X chromosome than on the second chromosome
in the data set of Caracristi and Schlötterer (2003) (data
not shown). The low degree of admixture on the X
chromosome seems to contrast the observation that most of
the genes involved in Z-ness are located on the autosomes.
Nevertheless, it has to be noted that QTL maps, such as the
one constructed by Hollocher et al. (1997b), are not
expected to reflect the different selection intensities on X
chromosomes and autosomes under natural conditions.
Ecological Selection
As a consequence of adaptation to one habitat fitness
can be reduced in another habitat, leading to a reduction in
gene flow between populations adapted to two different
habitats (Cooper and Lenski 2000; Hawthorne and Via
2001; Hendry 2001). Both theoretical and experimental
studies demonstrated that such an ecological adaptation
mediated reduction in gene flow could even result in
speciation (Rice and Hostert 1993; Gavrilets 1999;
Schluter 2001; Ogden and Thorpe 2002). Examples could
be thermal adaptation or changes in life cycle to adapt to
temperate climates. Based on the population history of D.
melanogaster it has previously been suggested that the
habitat expansion out of Africa was facilitated by selective
sweeps (Begun and Aquadro 1995; Kirby and Stephan
1996; Kauer et al. 2002). Recently, various studies have
provided independent evidence that the fixation of
beneficial mutations is more common than assumed under
the neutral theory of molecular evolution (Kimura 1983;
Bustamante et al. 2002; Fay, Wyckoff, and Wu 2002;
Smith and Eyre-Walker 2002). Furthermore, adaptive explanations were also suggested for phenotypic traits differingbetweenAfricanandnon-Africanpopulations(Vouidibio
et al. 1989; Takahashi et al. 2001). If non-African flies
Nonneutral Admixture in Drosophila melanogaster 1335
Appendix B
Assignment of Individuals to the ‘‘European’’ Cluster Using Different Locus Sets of the Third Chromosome
Locus Set
Individuals
3-1
3-2
3-3
3-4
3-5
Mean over Locus Sets
Complete Data Set
Reduced Data Seta
Mean European
Zh1
Zh13
Zh16
Zh18
Zh2
Zh20
Zh21
Zh23
Zh24
Zh25
Zh26
Zh28
Zh29
Zh32
Zh33
Zh34
Zh35
Zh36
Zh39
Zh40
Zh42
Zh43
Mean Zh
Zs11
Zs15
Zs2
Zs22
Zs24
Zs28
Zs29
Zs30
Zs37
Zs40
Zs49
Zs53
Zs6
Mean ZS
0.98
0.35
0.26
0.01
0.09
0.06
0.01
0.50
0.02
0.02
0.00
0.03
0.07
0.11
0.10
0.03
0.03
0.02
0.00
0.01
0.64
0.01
0.09
0.11
0.06
0.10
0.00
0.30
0.01
0.03
0.14
0.02
0.01
0.01
0.01
0.03
0.01
0.06
0.97
0.99
0.97
0.84
0.50
0.41
0.03
0.54
0.06
0.01
0.01
0.45
0.01
0.01
0.02
0.66
0.84
0.05
0.01
0.01
0.01
0.98
0.01
0.34
0.01
0.03
0.01
0.02
0.01
0.01
0.01
0.02
0.09
0.01
0.06
0.01
0.01
0.02
0.99
0.04
0.58
0.53
0.01
0.16
0.03
0.01
0.02
0.00
0.08
0.15
0.01
0.00
0.01
0.05
0.03
0.01
0.01
0.01
0.00
0.00
0.00
0.08
0.00
0.01
0.01
0.01
0.01
0.00
0.00
0.00
0.01
0.01
0.02
0.01
0.01
0.01
0.97
0.05
0.95
0.13
0.38
0.04
0.04
0.01
0.10
0.03
0.01
0.11
0.01
0.05
0.01
0.71
0.01
0.01
0.01
0.02
0.01
0.03
0.15
0.13
0.03
0.01
0.03
0.05
0.01
0.02
0.02
0.01
0.02
0.01
0.00
0.06
0.24
0.04
0.96
0.05
0.13
0.02
0.32
0.33
0.26
0.24
0.06
0.05
0.82
0.05
0.45
0.01
0.07
0.06
0.90
0.53
0.09
0.01
0.57
0.01
0.12
0.23
0.01
0.02
0.10
0.16
0.02
0.01
0.01
0.01
0.06
0.01
0.02
0.01
0.02
0.03
0.98
0.30
0.58
0.31
0.26
0.20
0.07
0.26
0.05
0.02
0.18
0.16
0.11
0.03
0.04
0.30
0.36
0.12
0.02
0.01
0.25
0.21
0.08
0.18
0.02
0.03
0.03
0.11
0.01
0.02
0.04
0.01
0.04
0.01
0.02
0.02
0.06
0.03
0.99
0.41
0.66
0.31
0.35
0.17
0.05
0.32
0.22
0.02
0.04
0.38
0.04
0.02
0.19
0.30
0.47
0.17
0.02
0.01
0.17
0.10
0.06
0.20
0.01
0.16
0.02
0.32
0.01
0.02
0.01
0.01
0.06
0.00
0.07
0.02
0.01
0.06
0.99
0.49
0.55
0.40
0.29
0.12
0.04
0.41
0.15
0.03
0.03
0.38
0.09
0.02
0.26
0.32
0.44
0.23
0.02
0.01
0.27
0.14
0.07
0.22
0.01
0.13
0.03
0.30
0.01
0.02
0.01
0.02
0.07
0.01
0.08
0.04
0.02
0.06
a
Data set with the same number of European and Zimbabwean chromosomes.
have accumulated many mutations enhancing fitness outside of Africa, they can be expected to have lost some
of their adaptations to the African environment. Such
ecological trade-offs (‘‘antagonistic pleiotropy’’) were
demonstrated in experimental E. coli populations (Cooper
and Lenski 2000; Cooper, Bennett and Lenski 2001).
European alleles could, therefore, be deleterious in an
African context and should be purged from the
population by natural selection. Our results also have
similarity to other studies, which described strong
selection against hybrids and thus a maintenance of the
integrity of divergent populations (Via, Bouck, and
Skillman 2000; Hawthorne and Via 2001). However we
have no direct evidence for antagonistic pleiotropy of
differential adapted alleles in and outside of Africa.
Therefore, it remains speculative whether our results
could be explained by environmental selection.
Stronger Selection on the X Chromosome
The observed difference between X chromosomes and
autosomes could be caused by hemizygosity of the X
chromosomes in males. As males carry only a single copy of
the X chromosomes, selection is more effective for recessive
mutations located on the X chromosome (Charlesworth,
Coyne, and Barton 1987; Aquadro, Begun, and Kindahl
1994). Assuming that African populations (and urban
populations in particular [see above]) are challenged by
a continuous influx of non-African alleles, and these alleles
are recessive and deleterious in Zimbabwean flies, selection
will remove non-African alleles more efficiently on the X
chromosome than on the autosomes. Thus, European alleles
are maintained for longer time spans on the autosomes,
leading to a higher proportion of European alleles on
the autosomes. Irrespective of the selective force acting,
hemizygosity of X chromosomes in males could therefore
explain the observed difference between X chromosomes
and autosomes.
Alternatively, if the number of European alleles
deleterious in Zimbabwe is higher on the X chromosome,
this could also explain the sheer absence of European alleles
on the Zimbabwean X chromosomes. This hypothesis
seems, however, unlikely given that X chromosomal gene
density is not higher than the autosomal gene density (Hey
1336 Kauer et al.
and Kliman 2002). Finally, a higher impact of selection on
the X chromosome could be due to stronger epistatic
interactions on the X chromosome.
Acknowledgments
We are thankful to R. Butlin, D. Falush, B. Harr, J.
Pool, M. Richie, C. Vogl, members of the CS lab, and two
anonymous reviewers for helpful comments on earlier
versions of the manuscript. Many thanks to C. Aquadro, B.
Harr, J. Gorczyca, and C. I. Wu for collecting and sharing
fly stocks. This work was supported by FWF grants to C.S.
Literature Cited
Alipaz, J. A., C. I. Wu, and T. L. Karr. 2001. Gametic incompatibilities between races of Drosophila melanogaster.
Proc. R. Soc. Lond. B Biol. Sci. 268:789–795.
Andolfatto, P. 2001. Contrasting patterns of X-linked and
autosomal nucleotide variation in Drosophila melanogaster
and Drosophila simulans. Mol. Biol. Evol. 18:279–290.
Aquadro, C. F., D. J. Begun, and E. C. Kindahl. 1994. Selection,
recombination, and DNA polymorphism in Drosophila. Pp.
46–56 in B. Golding, ed. Non-neutral evolution. Chapman &
Hall, London.
Begun, D. J., and C. F. Aquadro. 1992. Levels of naturally
occurring DNA polymorphism correlate with recombination
rates in D. melanogaster. Nature 356:519–520.
———. 1993. African and North American populations of
Drosophila melanogaster are very different at the DNA level.
Nature 356:548–550.
———. 1995. Evolution at the tip and base of the X
chromosome in an African population of Drosophila
melanogaster. Mol. Biol. Evol. 12:382–390.
Bustamante, C. D., R. Nielsen, S. A. Sawyer, K. M. Olson, M. D.
Purugganan, and D. L. Hartl. 2002. The cost of inbreeding in
Arabidopsis. Nature 416:531–534.
Capy, P., E. Pla, and J. R. David. 1993. Phenotypic and genetic
variability of morphometrical traits in natural populations of
Drosophila melanogaster and D. simulans. I. Geographic
variations. Genet. Sel. Evol. 25:517–536.
Capy, P., M. Veuille, M. Paillette, J. M. Jallon, J. Vouidibio, and
J. R. David. 2000. Sexual isolation of genetically differentiated sympatric populations of Drosophila melanogaster in
Brazzaville, Congo: the first step towards speciation? Heredity
84:468–475.
Caracristi, G., and C. Schlötterer. 2003. Genetic differentiation
between American and European D. melanogaster populations could be attributed to admixture of African alleles Mol.
Biol. Evol. 20:792–799.
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.
Cooper, V. S., A. F. Bennett, and R. E. Lenski. 2001. Evolution
of thermal dependence of growth rate of Escherichia coli
populations during 20,000 generations in a constant environment. Evolution 55:889–896.
Cooper, V. S., and R. E. Lenski. 2000. The population genetics
of ecological specialization in evolving Escherichia coli
populations. Nature 407:736–739.
Daborn, P., S. Boundy, J. Yen, B. Pittendrigh, and R. FrenchConstant. 2001. DDT resistance in Drosophila correlates with
cyp6g1 over-expression and confers cross-resistance to the
neonicotinoid imidacloprid. Mol. Genet. Genomics 266:
556–563.
David, J. R. and P. Capy. 1988. Genetic variation of Drosophila
melanogaster natural populations. Trends Genet. 4:106–111.
Dieringer, D., and C. Schlötterer. 2003. Microsatellite analyzer
(MSA)—a platform independent analysis tool for large
microsatellite data sets. Mol. Ecol. Notes 3:167–169.
Fay, J. C., G. J. Wyckoff, and C. I. Wu. 2002. Testing the neutral
theory of molecular evolution with genomic data from
Drosophila. Nature 415:1024–1026.
Gavrilets, S. 1999. A dynamical theory of speciation on holey
adaptive landscapes. Am. Nat. 154:1–22.
Hawthorne, D. J., and S. Via. 2001. Genetic linkage of ecological
specialization and reproductive isolation in pea aphids. Nature
412:904–907.
Hendry, A. P. 2001. Adaptive divergence and the evolution of
reproductive isolation in the wild: an empirical demonstration using introduced sockeye salmon. Genetica 113:515–
534.
Hey, J., and R. M. Kliman. 2002. Interactions between natural
selection, recombination and gene density in the genes of
Drosophila. Genetics 160:595–608.
Hollocher, H., C.-T. Ting, F. Pollack, and C.-I. Wu. 1997a.
Incipient speciation by sexual isolation in Drosophila
melanogaster: variation in mating preference and correlation
between sexes. Evolution 51:1175–1181.
Hollocher, H., C.-T. Ting, M. L. Wu, and C.-I. Wu. 1997b.
Incipient speciation by sexual isolation in Drosophila
melanogaster: extensive genetic divergence without reinforcement. Genetics 147:1191–1201.
Kauer, M., B. Zangerl, D. Dieringer, and C. Schlötterer. 2002.
Chromosomal patterns of microsatellite variability contrast
sharply in African and non-African populations of Drosophila
melanogaster. Genetics 160:247–256.
Kimura, M. 1983. The neutral theory of molecular evolution.
Cambridge University Press, Cambridge.
Kirby, D. A., and W. Stephan. 1996. Multi-locus selection and
the structure of variation at the white gene of Drosophila
melanogaster. Genetics 144:635–645.
Lemeunier, D., and S. Aulard. 1992. Inversion polymorphism in
Drosophila melanogaster. Pp. 339–405 in C. B. Krimbas and
J. R. Powell, eds. Drosophila inversion polymorphism. CRC
Press, Cleveland.
Maynard Smith, J., and J. Haigh. 1974. The hitch-hiking effect of
a favorable gene. Genet. Res. 23:23–35.
Ogden, R., and R. S. Thorpe. 2002. Molecular evidence for
ecological speciation in tropical habitats. Proc. Natl. Acad.
Sci. USA 7:7.
Pritchard, J. K., M. Stephens, and P. Donnelly. 2000. Inference of
population structure using multilocus genotype data. Genetics
155:945–959.
Rice, W., and E. Hostert. 1993. Laboratory experiments on
speciation: what have we learned in 40 years? Evolution 47:
1637–1653.
Schlötterer, C., and B. Harr. 2002. Single nucleotide polymorphisms derived from ancestral populations show no evidence
for biased diversity estimates in Drosophila melanogaster.
Mol. Ecol. 11:947–950.
Schlötterer, C., and B. Zangerl. 1999. The use of imperfect
microsatellites for DNA fingerprinting and population genetics. Pp. 153–165 in J. T. Epplen and T. Lubjuhn, eds.
DNA profiling and DNA fingerprinting. Birkhäuser, Basel,
Switzlerland.
Schluter, D. 2001. Ecology and the origin of species. Trends
Ecol. Evol. 16:372–380.
Smith, N. G., and A. Eyre-Walker. 2002. Adaptive protein
evolution in Drosophila. Nature 415:1022–1024.
Sokal, R. R., and F. J. Rohlf. 1995. Biometry. W. H. Freeman,
New York.
Nonneutral Admixture in Drosophila melanogaster 1337
Takahashi, A., S. C. Tsaur, J. A. Coyne, and C. I. Wu. 2001. The
nucleotide changes governing cuticular hydrocarbon variation
and their evolution in Drosophila melanogaster. Proc. Natl.
Acad. Sci. USA 98:3920–3925.
Ting, C. T., A. Takahashi, and C. I. Wu. 2001. Incipient
speciation by sexual isolation in Drosophila: concurrent
evolution at multiple loci. Proc. Natl. Acad. Sci. USA 98:
6709–6713.
Via, S., A. C. Bouck, and S. Skillman. 2000. Reproductive
isolation between divergent races of pea aphids on two hosts.
II. Selection against migrants and hybrids in the parental
environments. Evolution 54:1626–1637.
Vouidibio, J., P. Capy, D. Defaye, E. Pla, J. Sandrin, A. Csink,
and J. David. 1989. Short-range genetic structure of
Drosophila melanogaster populations in an Afrotropical
urban area and its significance. Proc. Natl. Acad. Sci. USA
86:8442–8446.
Weir, B. S., and C. C. Cockerham. 1984. Estimating F-statistics
for the analysis of population structure. Evolution 38:1358–
1370.
Wu, C.-I., H. Hollocher, D. J. Begun, C. F. Aquadro, Y. Xu, and
M.-L. Wu. 1995. Sexual isolation in Drosophila melanogaster: a possible case of incipient speciation. Proc. Natl. Acad.
Sci. USA 92:2519–2523.
David Rand, Associate Editor
Accepted April 4, 2003