Population structure of Glossina palpalis gambiensis (Diptera: Glossinidae)
between river basins in Burkina Faso: Consequences for area-wide
integrated pest management
Je ´re ´my Bouyera,b,*, Sophie Ravelc, Laure Guerrinid, Jean-Pierre Dujardine, Issa Sidibe ´f,
Marc J.B. Vreyseng, Philippe Solanoc,f, Thierry De Meeu ˆsc,f,h
aCirad, UMR Contro ˆle des maladies animales exotiques et e ´mergentes, Campus International de Baillarguet, F34398 Montpellier, France
bIsra-Lnerv, Service de Parasitologie, BP 2057 Dakar-Hann, Senegal
cInstitut de Recherche pour le De ´veloppement, Unite ´ mixte de Recherche IRD-CIRAD 177, Campus International de Baillarguet, 34398 Montpellier Cedex 5, France
dCirad, UMR AGIRs, Campus International de Baillarguet, F34398 Montpellier, France
eUR 165 UMR 2724 GEMI, Mahidol University, Bldg. 2, 999 Phuttamonthon 4Rd., Nakhon Pathom 73170, Thailand
fCentre International de Recherche-de ´veloppement sur l’Elevage en Zone Subhumide, BP 454 Bobo-Dioulasso, Burkina Faso
gEntomology Unit, FAO/IAEA Agriculture and Biotechnology Laboratory, Joint FAO/IAEA Programme of Nuclear Techniques in Food and Agriculture, A-2444 Seibersdorf, Austria
hCNRS, De ´le ´gation Languedoc-Roussillon, 1919, route de Mende, 34293 Montpellier Cedex 5, France
In Burkina Faso as in other parts of West Africa, African animal
trypanosomosis (AAT) is a major obstacle to the development of
more efficient and sustainable livestock production systems (Itard
et al., 2003). In 2001, African countries launched the Pan African
Tsetse and Trypanosomiasis Eradication Campaign (PATTEC),
which aims to increase efforts to manage this major plague. AAT
is indeed considered one of the root causes of hunger and poverty
in most sub-Saharan African countries where it represents a
serious impediment to sustainable agricultural rural development
Pattec.htm). In the sub-humid savannah of West Africa, riverine
tsetse species such as Glossina palpalis gambiensis Vanderplank
inhabit riparian forests along river systems where they are major
vectors of AAT (Buxton, 1955; Challier, 1973, 1982; Bouyer et al.,
2006; Guerrini et al., 2008) and human African trypanosomosis
(HAT) or sleeping sickness (Mulligan, 1970; Camara et al., 2006).
Control of tsetse can be achieved through a variety of
techniques (Cuisance et al., 2003), including traps, insecticide-
impregnated targets (Green, 1994), live-baits (Bauer et al., 1992;
Infection, Genetics and Evolution 10 (2010) 321–328
A R T I C L E I N F O
Received 4 November 2009
Received in revised form 20 December 2009
Accepted 25 December 2009
Available online 7 January 2010
Area-wide integrated pest management
A B S T R A C T
African animal trypanosomosis is a major obstacle to the development of more efficient and sustainable
livestock production systems in West Africa. Riverine tsetse species such as Glossina palpalis gambiensis
Vanderplank are their major vectors. A wide variety of control tactics is available to manage these
vectors, but their elimination will only be sustainable if control is exercised following area-wide
integrated pest management (AW-IPM) principles, i.e. the control effort is targeting an entire tsetse
population within a circumscribed area. In the present study, genetic variation at microsatellite DNA loci
was used to examine the population structure of G. p. gambiensis inhabiting two adjacent river basins, i.e.
the Comoe ´ and the Mouhoun River basins in Burkina Faso. A remote sensing analysis revealed that the
woodland savannah habitats between the river basins have remained unchanged during the last two
decades. In addition, genetic variation was studied in two populations that were separated by a man-
made lake originating from a dam built in 1991 on the Comoe ´. Low genetic differentiation was observed
between the samples from the Mouhoun and the Comoe ´ River basins and no differentiation was found
between the samples separated by the dam. The data presented indicate that the overall genetic
differentiation of G. p. gambiensis populations inhabiting two adjacent river basins in Burkina Faso is low
(FST= 0.016). The results of this study suggest that either G. p. gambiensis populations from the Mouhoun
are not isolated from those of the Comoe ´, or that the isolation is too recent to be detected. If elimination
of the G. p. gambiensis population from the Mouhoun River basin is the selected control strategy, re-
invasion from adjacent river basins may need to be prevented by establishing a buffer zone between the
Mouhoun and the other river basin(s).
? 2010 Elsevier B.V. All rights reserved.
* Corresponding author at: Isra-Lnerv, Service de Parasitologie, BP 2057 Dakar-
Hann, Se ´ne ´gal. Tel.: +221 33 951 02 65; fax: +221 33 832 36 79.
E-mail address: firstname.lastname@example.org (J. Bouyer).
Contents lists available at ScienceDirect
Infection, Genetics and Evolution
journal homepage: www.elsevier.com/locate/meegid
1567-1348/$ – see front matter ? 2010 Elsevier B.V. All rights reserved.
Bouyer et al., 2007b, 2009a), sequential aerial technique (Kgori
et al., 2006), and the sterile insect technique (SIT) (Vreysen et al.,
2000). In the past, most control efforts ended by a recovery of
tsetse populations due to either flies surviving the initial
interventions, or flies emigrating from untreated regions, or both
(Hargrove, 2003). The strategic choice between elimination and
suppression of a tsetse population is of prime importance as it will
have significant economic implications. In that respect, knowledge
of the structure of the target population can facilitate this critical
decision-making (Camara et al., 2006; Solano et al., 2009). For
isolated populations, tsetse elimination is undoubtedly the most
cost-effective strategy, as was demonstrated with the elimination
of Glossina austeni Newstead from the Island of Unguja, Zanzibar in
1994–1997 (Vreysen et al., 2000). On mainland Africa, the
geographical limits of the target tsetse populations are less clearly
defined, but analyses of the gene frequencies of the various
subpopulations can help understanding and quantifying gene
flow. This would facilitate the development of appropriate
control strategies and the decision-making on the need for a
buffer zone during and after area-wide elimination campaigns
In the Mouhoun river basin, contrarily to morsitans group flies
(Bouyer et al., 2005), riverine species of tsetse seem to be resilient
to man-made changes. This has been associated with their ability
to easily adapt to peridomestic situations and their linear habitat
that allows them to easily disperse between favorable patches, i.e.
riverine forests acting as ‘‘genetic corridors’’ (Cuisance et al., 1985;
Bouyer et al., 2007a), but also to more opportunistic host
preferences (Weitz, 1963). In Burkina Faso, G. p. gambiensis
populations inhabiting fragmented habitat of the agro-pastoral
area of Side ´radougou were found to be genetically structured
(Solano et al., 2000). A recent study also revealed a significant level
of structuring (FST= 0.012) but no complete genetic isolation
between populations inhabiting a 216-km section of the Mouhoun
river (Bouyer et al., 2007c); an area which is the target of a control
campaign launched by the Government of Burkina under the
African Union-PATTEC initiative (Fond Africain de De ´veloppement,
2004). A more precise investigation revealed a structure at a finer
scale with a significant isolation by distance between neighbor-
hoods not wider than 1 km (Bouyer et al., 2009b). The same study
suggested that across longer distances (i.e. along the 216 km of the
river section studied) the weak differentiation measured results
from a combination of local structure effects and a non-
equilibrium situation (recent fractioning).
was used to examine the population structure of G. p. gambiensis
between the adjacent Comoe ´ and Mouhoun river basins, and
between two sides of a dam. The objective was to assess tsetse
population structuring in different river basins, its relation to
natural barriers and its consequences for potential area-wide
integrated pest management(AW-IPM) campaigns (Klassen, 2005;
Vreysen et al., 2007).
2. Materials and methods
2.1. Study site
One site on the Mouhoun River (M) and two sites on the Comoe ´
River (C1 and C2) were sampled using 10 biconical traps in each
site (Challier and Laveissie `re, 1973) (Fig. 1) in June and July 2006.
Within each sample site, the maximal distance between traps was
500 m. From upstream to downstream, C1 is located before and C2
after a dam. In this area, the tributaries of the Mouhoun and Comoe ´
that were sampled are separated by only 2–5 km of woody
savannah. The between samples geographic distances are 13 km
between M and C1 and 6 km between C1 and C2.
2.2. Sampling and genotyping
A total of 30 G. p. gambiensis from site M (19 males and 11
females), 20 from site C1 (11 males and 9 females) and 40 from site
C2 (32 males and 8 females) were genotyped (see number of flies
genotyped by trapping site in Table 1) using 9 microsatellite loci:
Gpg 55.3 (Solano et al., 1997), B104, B11, B110, C102 (kindly
supplied by A. Robinson, Entomology Unit, Food and Agricultural
Organization of the United Nations/International Atomic Energy
Agency [FAO/IAEA], Agriculture and Biotechnology Laboratory,
Seibersdorf, Austria), GpCAG133 (Baker and Krafsur, 2001), pGp11,
pGp13, pGp24 (Luna et al., 2001). From these, B104, B110, pGp11,
pGp13, and 55.3 are known to be located on the X chromosome.
CAG133 have trinucleotide repeats whereas the others are
The legs of each individual tsetse were removed, transferred to
a tube to which 200 ml of 5% Chelex1chelating resin was added
(Walsh et al., 1991; Solano et al., 2000). After incubation at 56 8C
for 1 h, DNA was denatured at 95 8C for 30 min. The tubes were
then centrifuged at 12,000 ? g for two min and frozen for later
analysis. The PCR reactions were carried out in a thermocycler (MJ
Research, Cambridge, UK) in 10 ml final volume, using 1 ml of the
supernatant from the extraction step. After PCR amplification,
allele bands were routinely resolved on a 4300 DNA Analysis
System from LI-COR (Lincoln, NE) after migration in 96-lane
reloadable (3?) 6.5% denaturing polyacrylamide gels. This method
allows multiplexing by the use of two infrared dyes (IRDyeTM),
separated by 100 nm (700 and 800 nm), and read by a two channel
detection system that uses two separate lasers and detectors to
eliminate errors due to fluorescence overlap. To determine the
different allele sizes, a large panel of about 30 size markers was
used. These sizemarkers had been previouslygenerated by cloning
alleles from individual tsetse flies into pGEM-T Easy Vector
(Promega Corporation, Madison, WI, USA). Three clones of each
allele were sequenced using the T7 primer and the Big Dye
Terminator Cycle Sequencing Ready Reaction Kit (PE Applied
Biosystems, Foster City, CA, USA). Sequences were analysed on a PE
Applied Biosystems 310 automatic DNA sequencer (PE Applied
Biosystems) and the exact size of each cloned allele was
determined. PCR products from these cloned alleles were run in
Number of flies genotyped in each trapping site and location of these trapping sites
Site Trap numberLongitudeLatitudeNumber of flies
J. Bouyer et al./Infection, Genetics and Evolution 10 (2010) 321–328
the same acrylamide gel as the samples, allowing the allele size of
the samples to be determined accurately (Ravel et al., 2007). The
gels were read twice by two independent readers.
2.3. Remote sensing analysis
Two Landsat 7 enhanced thematic mapper plus (ETM+) images
from 16 November 1986 and 17 November 2001 (Path/row: 197-
52) were used to assess the development of the savannah area
between the two river basins. These scenes were cloud free and
hadno apparenthaze.Theriverpathsweredigitised from Landsat
images by using Mapinfo 7.0 software. A supervised classification
using a maximum likelihood classifier (ENVI 4.3 software) from a
three-channel composition (TM4, TM3, TM2) (Girard and Girard,
basins where the tsetse populations were sampled (Fig. 1)
(Guerrini et al., 2008). Five land-use classes were identified and
described using the standard nomenclature of African vegetation
types (Aubreville, 1957): water, swamp forests, woodland
savannah, crops and bare ground. A fragmentation analysis
included only one class of habitat, i.e. woodland savannah, which
is supposed to be the most favorable for tsetse dispersal. For both
1986 and 2001, the following common metrics were calculated
with Mapinfo 7.0 software: class area, corresponding to the total
area of suitable habitat; numberof patchesof suitable habitat and
mean patch size (Huang et al., 2006). These metrics are simple to
obtain over large areas and are likely to be correlated to the
fluidity of the environment to dispersal. For example, the
occurrence of corridors of suitable habitats between the river
basins is correlated to the patches density in the area between
these river basins. The mean patch sizes were compared between
the 2 years using the Wilcoxon rank sum test (Hollander and
Wolfe, 1973) under the R2.9.2 free statistical software (R
Development Core Team, 2008).
Fig. 1. Location of the three sampling sites of G. p. gambiensis on a tributary of the Mouhoun River basin (M) and the Comoe ´ River Basin (C1 and C2), Burkina Faso. The small
maps on the right present the areas of woodland savannah separating the two river basins in 1986 and 2001.
J. Bouyer et al./Infection, Genetics and Evolution 10 (2010) 321–328
2.4. Data analysis
Population structure was assessed through Wright’s F-statistics
(Wright, 1965). FISmeasures genetic identity between genes of
individuals relative to identity within subsamples and FST
measures genetic identity between genes of subsamples relative
to identity within the total sample. Hence, FISis a measure of
deviation from random union of gametes within subsamples and
FSTis a measure of deviation from random dispersal of individuals
between subpopulations, and thus of population differentiation,
providing subsamples indeed match actual subpopulations. F-
statistics were estimated with Weir and Cockerham’s unbiased
estimators (Weir and Cockerham, 1984) and tested by randomisa-
tion procedures implemented in Fstat 18.104.22.168 (Goudet, 2002,
updated from Goudet, 1995). To test for the significant deviation
from 0 of FIS, alleles were randomised within subsamples and the
statistic used was the estimator itself. For differentiation,
individuals were randomised between subsamples and the
statistic used was the log likelihood ratio G (Goudet et al.,
1996). Because some loci are X-linked and thus haploid in males,
we coded such loci as missing data for FIS analyses or as
homozygous for FSTanalyses in male individuals. When appropri-
ate, we also used 95% confidence intervals of bootstrap over loci or
jackknife over populations (when bootstrap were unavailable) as
given by Fstat.
Null allele presence and scoring errors were assessed with
Micro-Checker v 2.2.3. (Van Oosterhoutet al., 2004; Shipley,2005).
For each locus, null allele frequencies estimated with Brookfield’s
second method (Brookfield, 1996) were used to estimate the
expected frequency of blank homozygotes, which were compared
with the observed number of blanks at the considered locus. To
maximise the analysis the mean expected frequency of blanks was
compared to the sum of all blanks observed at a particular locus
across all subsamples with a binomial unilateral test (H1: there are
less blanks observed than expected). Binomial tests were
undertaken with R version 2.9.2 (2008) (R Development Core
Team, 2008). For these analyses, X-linked loci were analysed on
females only. The procedure used in Micro-Checker to detect short
allele dominance in each subsample independently is less
powerful as compared to the multiple regression approach (De
Meeu ˆs et al., 2004). For each locus, we thus analysed with R the
relationship between FIS, the corresponding size of the allele and
the subsample it belongs to.
Linkage disequilibrium between loci was evaluated between
each locus pair over all subsamples with the G-based test
implemented in Fstat. With nine loci this leads to 36 tests to be
handled. Multiple testing enhances the chance to get significant
tests (here about two significant tests are expected under the null
hypothesis at the 5% level of significance). We tested if the
observed number of significant tests at the 5% level is significantly
greater than the 5% expected with a binomial test with mean 0.05.
In order to detect which locus pairs, if any, are in significant
linkage, we also adjusted the P-values to the sequential Bonferroni
level (Holm, 1979) by multiplying the P-values by the number of
remaining tests (De Meeu ˆs et al., 2007).
Sex biased dispersal bilateral tests were undertaken using Fstat
22.214.171.124 with the FSTbased test and the mean assignment index
(multilocus probability to belong to the sampling site) corrected
for population effects and its variance (mAIcand vAIc) (Favre et al.,
1997) as recommended in Goudet et al. (2002). We used 10,000
permutations of sex status of individuals within samples (Goudet
et al., 2002) and applied the tests within each site (C1, C2 and M)
considering each trap containing at least one female and one male
as a subpopulation.
Within each site, the GPS coordinates of each trap were
available. This allowed implementing isolation by distance
between individuals test (Rousset, 2000; Leblois et al., 2003;
Watts et al., 2007). This test uses an equivalent of the statistic FST/
(1 ? FST) (Rousset, 1997), either a ˆ (Rousset, 2000) or e ˆ (Watts et al.,
2007), that is connected to the distance between individuals dInd
with the equation: a ˆ (or e ˆ) = (1/4Ds2)dInd+ Constant, where D is
the density of reproducing individuals and s the average distance
between reproducers and their parents (Rousset, 1997, 2000;
Leblois et al., 2003; Watts et al., 2007) in a one-dimensional
framework as it is the case here. Hence, the slope of the regression
b between genetic and geographical distance, with an independent
knowledge of density, should allow estimating the mean dispersal
of individuals as s ¼
ship is assessed through a Mantel test (Mantel, 1967) using both
statistics a ˆ and e ˆas recommended (Watts et al., 2007). The statistic
e ˆ is more powerful but biased as compared to a ˆ. In particular,
estimates are more accurate with a ˆ when e ˆ leads to 4Ds2< 10,000
(Watts et al., 2007). In such case the slope b and its 95% Bootstrap
confidence intervals (CI) were estimated using a ˆ as recommended
(Watts et al., 2007).
Effective subpopulation sizes were estimated in each subsam-
ple containing at least four genotyped tsetse flies. We used and
compared different Neestimation methods: the linkage disequi-
(2008) (LDWD), and the method (Estim) of Vitalis and Couvet
(2001a,b,c). Bartley’s method was implemented with NeEstimator
Do, 2008) and Vitalis and Couvet’s was implemented by Estim 1.2
(Vitalis and Couvet, 2001a).
More than three levels (i.e. individuals, subpopulations and
total) exist in the tsetse sample. Individuals were caught in
different traps along rivers, in two different river basins where
subdivision might exist between sites from both sides of the dam
(C1 and C2).HIERFSTAT version 0.03-2 (Goudet, 2005) is a packagefor
the statistical software R. This package computes hierarchical F-
statistics from any number of hierarchical levels (Goudet, 2005).
FTrap/Site represents the homozygosity due to subdivision into
different traps within sites and was tested by randomizing
individuals between traps within the same site. FC1C2/River
measures the relative homozygosity due to the geographical
separation between sites C1 and C2 by dams and was tested by
randomizingtraps (withallindividuals contained)between C1and
C2. FRiver/Total measures the relative homozygosity due to the
subdivision into different river basins and was tested by
randomizing traps across river basins (because dam was not
found to be a significant barrier). These tests were performed with
HIERFSTAT. A user-friendly step-by-step tutorial of how using
HierFstat can be found in (De Meeu ˆs and Goudet, 2007).
To get an overall picture of genotypic distribution across river
basins, dams and traps, a NJTREE was computed by the MEGA 3.1
software (Kumar et al. 2005, updated from Kumar et al., 2004). As
recommended (e.g. Takezaki and Nei, 1996; De Meeu ˆs et al., 2007),
the unrooted tree was built out of a Cavalli-Sforza and Edwards
chord distance matrix (Cavalli-Sforza and Edwards, 1967) with
Genetix4.03 (GENETIX,logicielsous WindowsTMpourla ge ´ne ´tique
des populations. Laboratoire Ge ´nome, Populations, Interactions
CNRS UMR 5000, Universite ´ de Montpellier II, Montpellier,France).
. The significance of the relation-
3.1. Remote sensing analysis
The remote sensing analysis showed that the total area of
woodland savannah was very similar in 1986 and 2001: woodland
savannah covered 58% and 63% of the total area in 1986 and 2001,
asdemonstrated bya patch numberof 4597and 4507,respectively
(Fig. 1). The mean patch size was also similar, i.e. on average 0.038
J. Bouyer et al./Infection, Genetics and Evolution 10 (2010) 321–328
and 0.042 km2in 1986 and 2001, respectively (Wilcoxon rank sum
test, W = 10423888, P-value = 0.607), providing similar dispersal
conditions during the two periods.
3.2. Within individuals genetic structure within subsamples
There was a strong and significant heterozygote deficit
(FIS= 0.183, P < 0.0001). This result was extremely variable across
sites as shown in Fig. 2. It appeared thus that an analysis of
differentiation between traps was required. Considering traps as
population units weakly improved the FISestimates in C1 and M
but not in C2. The FSTanalysis resulted in the following values:
FST= 0.087 (0.022–0.163 = 95% bootstrap confidence interval)
(P = 0.626),
FST= 0.041 (?0.006 to 0.081) (P = 0.049) for C1, C2 and M,
respectively. It thus seems that C1 and may be M is composed
of several units that correspond more or less to the different traps,
and C2 corresponds to a single reproductive unit. Consequently,
traps will be considered separately for C1 and M while C2 will be
considered as a single unit.
Using this sampling design, FIS= 0.159 still presents a strong
and highly significant heterozygote deficit (P < 0.0001). As shown
in Fig. 3, this heterozygote deficit appears extremely variable
(?0.053 to 0.046)(P = 0.479) and
across lociasexpectedinpresence ofnull allelesand/or short allele
dominance. Microchecker analysis lead to the conclusion that two
loci showed the presence of null alleles (B11 and pGp24), with a
very nice fit between expected and observed number of blanks (10
and 9.8 for B11, 16 and 15.6 for pGp24 for 72 observations, P = 0.61
in both cases).
A negative and marginally significant relationship between
allele size and FISwas found for only two loci (B11 and XB110). The
lack of power of the tests is related to the small sample size of most
subsamples, hence a huge variance in FISestimates. From Fig. 3, the
existence of short allele dominance might explain the huge FIS
variance for XB110.
Among the 36 pairs of loci, four appeared in significant linkage
which is not significantly different from the 5% expected
(P = 0.104). Nevertheless, two pairs appeared significantly linked
after sequential Bonferroni correction, both involving pGp24 (with
XpGp13 and C102).
Because we want the most possible precise estimation of
migration and subpopulation sizes, all these observations led us to
remove B11, XB110 and pGp24 from further analyses. As shown in
Fig. 3, this leads to a non-significant FISover all subsamples and
3.3. Differentiation between traps within sites with the six
Comoe ´ samples tend to show a female biased dispersal (female
Thus the signal is far from clear and, with only a single slightly
significant value in C2 for vAIc it is probably safer to assume that no
strong signal exists if any. Consequently, females and males were
pooled in the further analyses. Differentiation between traps was
smallornon-existentinComoesamples.FST= 0.053(P = 0.34)witha
95% CI = (?0.006, 0.119) in C1 and FST= 0 (P = 0.94) with a 95%
CI = (?0.079, 0.031) in C2. In the Mouhoun samples, differentiation
appeared slightly higher with a FST= 0.061 and a 95% CI = (0.017,
0.101), though marginally not significant (P = 0.06).
Isolation by distance tests were all non-significant except in the
Mouhoun sample with e ˆstatistic (P = 0.02) with 4Ds2= 8250 (with
the testismarginally significant,
4Ds2= 16767). We anyway know that such isolation by distance
does exist along the Mouhoun river with a 4Ds2= [3105, 39936]
(Bouyer et al., 2009a,b).
P-value = 0.09, with
Fig. 3. FISat the nine different loci with 95% confidence intervals (CI) obtained after jackknife over subsamples (different traps for C1 and M, and C2). The mean over loci with
95% CI of bootstrap over loci is also shown. Corresponding P-values are presented between brackets. Over all* was computed after removal of problematic loci (B11, XB110
and pGp24) (see text for details).
Fig. 2. FISestimates in the different sites with 95% confidence intervals obtained by
bootstrap over loci. Values when considering each site or each trap as a single
population are both given with the corresponding P-value between brackets.
J. Bouyer et al./Infection, Genetics and Evolution 10 (2010) 321–328
Consequently, differentiation is not confirmed within C2, but
this may be due to missing data (very few complete genotypes
were available for permutations), the total absence of differentia-
tion is confirmed in C1 and an increasing differentiation with
distance is confirmed along the Mouhoun river. It will thus be
wiser to consider each site from the Comoe (C1 and C2) as two
population units and keep tsetse flies from each trap along the
Mouhoun (M) as belonging to discrete entities.
3.4. Effective population sizes
Effective population sizes are presented in Table 2. When
several methods provide a real number (i.e. different from infinity)
values and confidence intervals are in the same order of
magnitude. Simulations undertaken earlier indicate that these
estimations are likely to be roughly accurate with the kind of
sampling undertaken (Bouyer et al., 2009b), but these values
should not be considered as accurate estimates. Some values may
appear small, may be as a result of a residual Wahlund effect (all
available methods are linkage disequilibrium based) and along the
Mouhoun where a perfect match between tsetse flies in one trap
and a corresponding subpopulation (or neighborhood) unit is not
3.5. Differentiation between sites
A HierFstat exploration of data show a positive differentiation
between traps (in Mouhoun basin) (FTrap/Site= 0.034), a negative
value between C1 and C2 (FC1C2/River= ?0.037) and a positive value
between river basins (FRiver/Total= 0.016), all these values being
non-significant (P-value > 0.21). If this latter value is corrected for
polymorphism (Hedrick, 2005), with a Hs? 0.7 this gives a
corrected F0? 0.05, which is still weak. The NJTREE unrooted tree
from Fig. 4 illustrates how little differentiation may occur between
the two river basins.
The aim of this study was to assess the level of isolation of G. p.
gambiensis populations between the Mouhoun and Comoe ´ River
basins, and within the Comoe ´ river following the construction of a
dam in 1991. This question has very important implications
because the National PATTEC program in Burkina Faso aims to
eliminate tsetse (G. p. gambiensis and G. tachinoides Westwood) in
the Mouhoun Riverbasin using an AW–IPM strategy (FondAfricain
de De ´veloppement, 2004), i.e. targeting an entire pest population
within a circumscribed area. Knowledge of the level of gene flow
between tsetse populations inhabiting different river basins will
facilitatethe decision-makingonthe need toestablishbuffer zones
between two river basins.
Null alleles are partly responsible for the overall positive Fis
value, as reported in other studies on G. palpalis (Camara et al.,
2006; Bouyer et al., 2007c; Ravel et al., 2007). A Wahlund effect,
caused by the presence of individuals originating from different
as reported in other studies in G. p. gambiensis (Solano et al., 2000)
and in G. p. palpalis (Ravel et al., 2007). Along the Mouhoun river,
where genetic isolation is linearly correlated with distance, the
distance between traps in our study was small (between 63 and
205 m) and a Wahlund effect is not expected to be strong. The
absence of significant isolation by distance in the Comoe ´ samples
might be related to the presence of more abundant blank
genotypes in these samples than in the Mouhoun sample, or
alternatively to a more important fragmentation observed along
the Mouhoun river, which tend to reduce tsetse dispersal (Bouyer
was reduced as much as possible in the present study (maximum
0.8 km) in comparison to previous studies (more than 10 km), in
order to reduce intra-sample clustering which lowers our ability to
detect isolation by distance, but is better for inter-populations
comparisons. Our results suggest that the sampling distance
should be even more reduced to avoid clustering.
We found low genetic differentiation between the samples
from the Mouhoun and the Comoe ´ river basins (0.016) and no
differentiation between the samples separated by the dam.
Dispersal of G. p. gambiensis around the man-made lake should
be reduced since the gallery forest has been completely flooded
and the savannah around the lake allow only two-dimensional
diffusion which reduced long distance movements in comparison
to one-dimensional diffusion. Moreover, the absence of boats on
this lake excludes passive transport of tsetse. However, the dam
built 17 years ago, has been unable to limit gene flow between the
Comoe ´ populations, at least at a detectable level. Actually, several
simulations were undertaken with Easypop 2.0.1 (Balloux, 2001)
using equation (5.112) in (Nei, 1975), the same number of
generations (approximately 100) and a mutation rate of u = 0.001,
which provided comparable FSTthan that observed for population
sizes of at least 1000 reproductive individuals. Since such
population sizes are not unrealistic, it is possible that gene flow
is negligible between C1 and C2 but that populations are too far
from equilibrium for any signature of a total separation to be
observed in Mali using mitochondrial DNA, i.e. substantial rates of
gene flow were detected among G. p. gambiensis populations
sampled along tributaries of three adjacent river basins (Marquez
et al., 2004).
Fig. 4. Unrooted NJTREE based on Cavalli-Sforzaand Edwards chord distancematrix
illustrating the population genetic structure of G. p. gambiensis across the two river
basins of the Comoe ´ (C1 and C2) and the Mouhoun (M) rivers, across sites C1 and C2
(separated by dams) and across traps in the Mouhoun (M with different numbers).
Effective population sizes in the different subsamples of size of at least four
genotyped tsetse individuals, with the three methods as described in the text. 95%
confidence intervals are presented between brackets.
52 (18, 1)
377 (46, 1)
2 (1, 3)
3 (2, 13)
23 (5, 1)
31 (11, 1)
6 (1, 1)
J. Bouyer et al./Infection, Genetics and Evolution 10 (2010) 321–328
Our data on gene frequencies of the G. p. gambiensis
populations in adjacent river basins are in concurrence with
data on the dispersal capacity of this species. During a mark-
release-recapture trial implemented along the Dienkoa river
near the sampling site of M, G. p. gambiensis flies were caught at
a distance of 2 km from the river during the rainy season
(Cuisance et al., 1983, 1985). This indicated that G. p. gambiensis
was able to cross 2–5 km of woodland savannah situated
between two river basins. It therefore appears that a fly, that has
reached this distance under the influence of increased relative
humidity in the savannah between the river systems, will
attempt to return to a gallery forest following an hygrometric
gradient after the rains and will have good or even equal
probabilities of joining the other river basin (Bouyer, 2006).
These data are in accordance with observations made in Mali
where during a release–recapture trial, laboratory-reared and
gamma-sterilised G. p. gambiensis males released in a tributary
of the Senegal River Basin were recaptured in a tributary of the
adjacent Niger River Basin (Z. Koudougou, K. Saleh, unpublished
reports to the IAEA).
The unstructured nature of the two G. p. gambiensis populations
studied seems to imply that the Mouhoun populations are not
isolated from those of the Comoe ´ basin, or that this isolation has
been a recent one, which is not very likely in view that the habitat
between the two river basins has remained unchanged for the last
two decades. It is acknowledged that the results might have been
influenced by the limited power of the tests used, or from a
sampling bias. This area is also used as a pastoral area by livestock
owners and the continuous movement of livestock could have
contributedto the passive dispersalof tsetse flies from onebasin to
If elimination of the G. p. gambiensis population of the
Mouhoun River basin is the selected control strategy, our results
indicate that re-invasion from adjacent river basins is probable
to occur and needs to be prevented by the establishment of
buffer areas between the Mouhoun and the Comoe ´. The
efficiency of such buffer areas will probably be higher when
the river basins are separated by landscapes unsuitable for tsetse
survival, like cotton crops, where significant amounts of
insecticides are used.
The target area of the present study was precisely selected
because the woody savannah between the two river basins was
very well conserved and consequently, it was hypothesised that
tsetsedispersal between the tworiver basinswas likelyto occur. In
other areas where agricultural activity is important between two
river basins, where the habitat is less preserved, or where the
distance between the end points of the tributaries of two river
basins is larger, dispersal of riverine tsetse between basins might
be much more difficult. This study needs to be expanded to all
West-African river basins from Senegal to Burkina Faso (Gambia,
Senegal, Niger, Comoe ´ and Mouhoun basins), to give a more
comprehensive analysis of the structure of riverine tsetse
populations at smaller scales. A stratified sampling scheme for
genetic studies will be developed and implemented using a
quantitative variable (the nearest distance between tributaries of
adjacent river basins), and a categorical one (the type of vegetation
between these tributaries).
This work was carried out with the financial support of IAEA
research contract n8 13303. Special thanks are due to the director
working conditions, to Fe ´lix Sanou, Wilfried Yoni, Bila Cene and
Adama Sana for their assistance during the field studies and
Damien Herault for genotyping the populations.
Aubreville, A., 1957. Accord a ` Yangambi sur la nomenclature des types africains de
ve ´ge ´tation. Bois For. Trop. 51, 23–27.
Baker, M.D., Krafsur, E.S., 2001. Identification and properties of microsatellite
markers in tsetse flies Glossina morsitans sensus lato (Diptera: Glossinidae).
Mol. Ecol. Notes 1, 234–236.
Balloux, F., 2001. EASYPOP (version 1.7): a computer program for population
genetics simulations. J. Hered. 92, 301–302.
Bartley, D., Bagley, M., Gall, G., Bentley, B., 1992. Use of linkage disequilibrium data
to estimate effective size of hatchery and natural fish populations. Cons. Biol. 6,
Bauer, B., Kabore, I., Liebisch, A., Meyer, F., Petrich-Bauer, J., 1992. Simultaneous
control of ticks and tsetse flies in Satiri, Burkina Faso, by the use of flumethrin
pour on for cattle. Trop. Med. Parasitol. 43, 41–46.
Bouyer, J., 2006. Ecologie des glossines du Mouhoun au Burkina Faso: inte ´re ˆt pour
l’e ´pide ´miologie et le contro ˆle des trypanosomoses africaines, Parasitologie.
Universite ´ Montpellier II, Montpellier, France, p. 206.
Bouyer, J., Guerrini, L., Ce ´sar, J., de la Rocque, S., Cuisance, D., 2005. A phyto-
sociological analysis of the distribution of riverine tsetse flies in Burkina Faso.
Med. Vet. Entomol. 19, 372–378.
Bouyer, J., Guerrini, L., Desquesnes, M., de la Rocque, S., Cuisance, D., 2006. Mapping
African animal trypanosomosis risk from the sky. Vet. Res. 37, 633–645.
diffusion of Glossina palpalis gambiensis (Diptera: Glossinidae) in Burkina Faso.
In: Vreysen, M.J.B., Robinson, A.S., Hendrichs, J. (Eds.), Area-wide Control of
Insect Pests. From Research to Field Implementation. Springer, Dordrecht, The
Netherlands, pp. 221–228.
Bouyer, J., Stachurski, F., Kabore ´, I., Bauer, B., Lancelot, R., 2007b. Tsetse control in
cattle from pyrethroid footbaths. Prev. Vet. Med. 78, 223–238.
Bouyer, J., Ravel, S., Vial, L., The ´venon, S., Dujardin, J.-P., de Meeus, T., Guerrini, L.,
(Diptera: Glossinidae) according to landscape fragmentation in the Mouhoun
river, Burkina Faso. J. Med. Entomol. 44, 788–795.
Bouyer, J., Stachurski, F., Gouro, A., Lancelot, R., 2009a. Control of bovine trypano-
somosis by restricted application of insecticides to cattle using footbaths. Vet.
Parasitol. 161, 187–193.
Bouyer, J., Balenghien, T., Ravel, S., Vial, L., Sidibe ´, I., The ´venon, S., Solano, P., De
Meeu ˆs, T., 2009b. Population sizes and dispersal pattern of tsetse flies: rolling
on the river? Mol. Ecol. 18, 2787–2797.
Brookfield, J.F.Y., 1996. A simple new method for estimating null allele frequency
from heterozygote deficiency. Mol. Ecol. 5, 453–455.
Buxton, P.A., 1955. The Natural History of Tsetse Flies. An Account of the Biology of
the Genus Glossina (Diptera) Lewis H.K. & Co. Ltd., London.
Camara, M., Harling Caro-Rian ˜o, H., Ravel, S., Dujardin, J.-P., Hervouet, J.-P., de
Meeu ¨s, T., Kagbadouno, M.S., Bouyer, J., Solano, P., 2006. Genetic and morpho-
metric evidence for isolation of a tsetse (Diptera: Glossinidae) population (Loos
islands, Guinea). J. Med. Entomol. 43, 853–860.
Cavalli-Sforza, L.L., Edwards, A.W.F., 1967. Phylogenetic analysis: model and esti-
mation procedures. Am. J. Hum. Genet. 19, 233–257.
Challier, A., 1973. Ecologie de Glossina palpalis gambiensis Vanderplank, 1949
(Diptera-Muscidae) en savane d’Afrique occidentale. ORSTOM, Paris.
Challier, A., 1982. The ecology of tsetse (Glossina spp.)(Diptera, Glossinidae): a
review. Insect. Sci. Appl. 3, 97–143.
Challier, A., Laveissie `re, C., 1973. Un nouveau pie `ge pour la capture des glossines
(Glossina: Diptera, Muscidae): description et essais sur le terrain. Cah. ORSTOM,
se ´r. Ent. Me ´d. Parasitol. 10, 251–262.
Cuisance, D., Fe ´vrier, J., Filledier, J., Dejardin, J., 1983. Etude sur le pouvoir de
dispersion des glossines. CRTA/IEMVT/OMS, Bobo-Dioulasso.
Cuisance, D., Fe ´vrier, J., Dejardin, J., Filledier, J., 1985. Dispersion line ´aire de Glossina
palpalis gambiensis et G. tachinoides dans une galerie forestie `re en zone sou-
dano-guine ´enne (Burkina Faso). Rev. Elev. Me ´d. ve ´t. Pays trop. 38, 153–172.
Cuisance, D., Itard, J., Solano, P., Desquesnes, M., Fre ´zil, J.L., Authie ´, E., 2003.
Trypanosomoses. Me ´thodes de lutte. In: Tec et Doc, Editions Me ´dicales Inter-
nationales, (Eds.), Principales maladies infectieuses et parasitaires du be ´tail.
Europe et Re ´gions chaudes, Lavoisier, Paris, France, pp. 139–165.
De Meeu ˆs, T., Goudet, J., 2007. A step-by-step tutorial to use HierFstat to analyse
populations hierarchically structured at multiple levels. Inf. Gen. Evol. 7, 731–
De Meeu ˆs, T., Humair, P.F., Grunau, C., Delaye, C., Renaud, F., 2004. Non-Mendelian
transmission of alleles at microsatellite loci: an example in Ixodes ricinus, the
vector of Lyme disease. Int. J. Parasitol. 34, 943–950.
De Meeu ˆs, T., McCoy, K.D., Prugnolle, F., Chevillon, C., Durand, P., Hurtrez-Bousse `s,
S., Renaud, F., 2007. Population genetics and molecular epidemiology or how to
‘‘de ´busquer la be ˆte’’. Inf. Gen. Evol. 7, 308–332.
Favre, L., Balloux, F., Goudet, J., Perrin, N., 1997. Female-biased dispersal in the
monogamous mammal Crocidura russula: evidence from field data and micro-
satellite patterns. Proc. R. Soc. B 264, 127–132.
Fond Africain de De ´veloppement, 2004. Projet multinational de cre ´ation de zones
durablement libe ´re ´es de la mouche tse ´-tse ´ et de la trypanosomiase dans six
pays sub-sahariens. De ´partement de l’Agriculure et du De ´veloppement Rural
(OCAR et ONAR), Tunis.
Girard, M.C., Girard, C.M., 1999. Traitement des donne ´es de te ´le ´de ´tection. Dunod,
Goudet, J., 1995. FSTAT (v. 1.2): a computer program to calculate F-statistics. J.
Hered. 86, 485–486.
J. Bouyer et al./Infection, Genetics and Evolution 10 (2010) 321–328
Goudet, J., 2005. HIERFSTAT, a package for R to compute and test hierarchical F- Download full-text
statistics. Mol. Ecol. Notes 5, 184–186.
Goudet, J., Raymond, M., De Meeu ˆs, T., Rousset, F., 1996. Testing differentiation in
diploid populations. Genetics 144, 1933–1940.
Goudet, J., Perrin, N., Waser, P., 2002. Tests for sex-biased dispersal using bi-
parentally inherited genetic markers. Mol. Ecol. 11, 1103–1114.
Green, C.H., 1994. Bait methods for tsetse fly control. Adv. Parasitol. 34, 229–291.
Guerrini, L., Bord, J.P., Ducheyne, E., Bouyer, J., 2008. Fragmentation analysis for
prediction of suitable habitat for vectors: the example of riverine tsetse flies in
Burkina Faso. J. Med. Entomol. 45, 1180–1186.
Hargrove, J.W., 2003. Tsetse Eradication: Sufficiency, Necessity and Desirability.
Centre for Tropical Veterinary Medicine, University of Edinburgh.
Hedrick, P.W., 2005. A standardized genetic differentiation measure. Evolution 59,
Hollander, M., Wolfe, D.A., 1973. Non Parametric Statistical Inference. John Wiley &
Sons, New York.
Holm, S., 1979. A simple sequentially rejective multiple test procedure. Scand. J.
Stat. 6, 65–70.
Huang, C., Geiger, E.L., Kupfer, J.A., 2006. Sensitivity of landscape metrics to
classification scheme. Int. J. Remote Sens. 27, 2927–2948.
Itard, J., Cuisance, D., Tacher, G., 2003. Trypanosomoses: Historique - Re ´partition
ge ´ographique. In: Lefe `vre, P.-C., Blancou, J., Chermette, R. (Eds.), Principales
maladies infectieuses et parasitaires du be ´tail. Europe et Re ´gions chaudes,
Lavoisier, Paris, pp. 1607–1615.
Kgori, P.M., Modo, S., Torr, S.J., 2006. The use of aerial spraying to eliminate tsetse
from the Okavango Delta of Botswana. Acta Trop. 99, 184–199.
Klassen, W., 2005. Area-wide integrated pest management and the sterile insect
technique. In: Dyck, V.A., Hendrichs, J., Robinson, A.S. (Eds.), Sterile insect
technique. Principles and practice in area-wide integrated pest management.
Springer, Dordrecht, The Netherlands, pp. 39–68.
Krafsur, E.S., 2003. Tsetse fly population genetics: an indirect approach to dispersal.
Trends Parasitol. 19, 162–166.
Kumar, S., Tamura, K., Nei, M., 2004. MEGA3: Integrated software for molecular
evolutionary genetics analysis and sequence alignment. Brief Bioinform. 5,
Leblois, R., Estoup, A., Rousset, F., 2003. Influence of mutational and sampling
factors on the estimation of demographic parameters in a ‘‘Continuous’’ popu-
lation under isolation by distance. Mol. Biol. Evol. 20, 491–502.
Luna, C., Bonizzoni, M., Cheng, Q., Robinson, A.S., Aksoy, S., Zheng, L., 2001.
Microsatellite polymorphism in tsetse flies (Diptera: Glossinidae). J. Med.
Entomol. 38, 376–381.
Mantel, N., 1967. The detection of disease clustering and a generalized regression
approach. Cancer Res. 27, 209–220.
Marquez, J.G., Vreysen, M.J.B., Robinson, A.S., Bado, S., Krafsur, E.S., 2004. Mitochon-
drial diversity analysis of Glossina palpalis gambiensis from Mali and Senegal.
Med. Vet. Entomol. 18, 1–8.
Mulligan, H.W., 1970. The African trypanosomes. G. Allen & Unwin LTD, London.
Nei, M., 1975. Molecular Population Genetics and Evolution. North-Holland Pub-
lishing Company, Amsterdam.
Peel, D., Ovenden, J.R., Peel, S.L., 2004. NeEstimator Version 1.3: Software for
Estimating Effective Population Size Computer Program. Queensland Govern-
ment, Department of Primary Industries and Fisheries.
Ravel, S., de Meeus, T., Dujardin, J.P., Ze ´ze ´, D.G., Gooding, R.H., Dusfour, I., Sane ´, B.,
Cuny, G., Solano, P., 2007. The tsetse fly Glossina palpalis palpalis is composed of
several genetically differentiated small populations in the sleeping sickness
focus of Bonon, Co ˆte d’Ivoire. Inf. Gen. Evol. 7, 116–125.
Development Core Team, R., 2008. R: A language and environment for statistical
computing computer program. R Foundation for Statistical Computing, Vienna
Rousset, F., 1997. Genetic differentiation and estimation of gene flow from F-
statistics under isolation by distance. Genetics 145, 1219–1228.
Rousset, F.,2000. Genetic differentiation betweenindividuals. J. Evol. Biol. 13, 58–62.
Shipley, P., 2005. Micro-Checker, Microsatellite Data Checking Computer Program.
University of Hull.
Solano, P., Duvallet, G., Dumas, V., Cuisance, D., Cuny, G., 1997. Microsatellite
markers for genetic population studies in Glossina palpalis (Diptera: Glossini-
dae). Acta Trop. 65, 175–180.
Solano, P., de La Rocque, S., de Me ´eus, T., Cuny, G., Duvallet, G., Cuisance, D., 2000.
Glossina palpalis gambiensis in the agropastoral zone of Sideradougou, Burkina
Faso. Insect Mol. Biol. 9, 433–439.
Solano, P., Ravel, S., Bouyer, J., Camara, M., Kagbadouno, M.S., Dyer, N., Gardes, L.,
Herault, D., Donnelly, M.J., De Meeu ˆs, T., 2009. The population structure of
Glossina palpalis gambiensis from island and continental locations in coastal
Guinea. PloS Negl. Trop. Dis. 3, e392.
Takezaki, N., Nei, M., 1996. Genetic distances and reconstruction of phylogenetic
trees from microsatellite DNA. Genetics 144, 389–399.
Van Oosterhout, C., Hutchinson, W.F., Wills, D.P.M., Shipley, P., 2004. MICRO-
CHECKER: software for identifying and correcting genotyping errors in micro-
satellite data. Mol. Ecol. Notes 4, 535–538.
Vitalis, R., Couvet, D., 2001a. ESTIM 1.0: a computer program to infer population
parameters from one- and two-locus gene identity probabilities. Mol. Ecol.
Notes 1, 354–356.
Vitalis, R., Couvet, D., 2001b. Two-locus identity probabilities and identity disequi-
librium in a partially selfing subdivided population. Genet. Res. 77, 67–81.
Vitalis, R., Couvet, D., 2001c. Estimation of effective population size and migration
rate from one- and two-locus identity measures. Genetics 157, 911–925.
Vreysen, M.J.B., Saleh, K.M., Ali, M.Y., Abdulla, A.M., Zhu, Z.-R., Juma, K.G., Dyck, V.A.,
Msangi, A.R., Mkonyi, P.A., Feldmann, H.U., 2000. Glossina austeni (Diptera:
Glossinidae) eradicated on the Island of Unguja, Zanzibar, using the sterile
insect technique. J. Econ. Entomol. 93, 123–135.
Vreysen, M., Robinson, A.S., Hendrichs, J., 2007. Area-wide Control of Insect Pests,
From Research to Field Implementation. Springer, Dordrecht, The Netherlands.
Walsh, P.S., Metzger, D.A., Higuchi, R., 1991. Chelex1100 as a medium for simple
extraction of DNA for PCR-based typing from forensic material. Biotechniques
Waples, R.S., Do, C., 2008. LDNE: a program for estimating effective population size
from data on linkage disequilibrium. Mol. Ecol. Res. 8, 753–756.
Watts, P.C., Rousset, F., Saccheri, I.J., Leblois, R., Kemp, S.J., Thompson, D.J., 2007.
Compatible genetic and ecological estimates of dispersal rates in insect (Coe-
nagrion mercuriale: Odonata: Zygoptera) populations: analysis of ‘neighbour-
hood size’ using a more precise estimator. Mol. Ecol. 16, 737–751.
Weir, B.S., Cockerham, C.C., 1984. Estimating F-statistics for the analysis of popula-
tions structure. Evolution 38, 1358–1370.
Weitz, B., 1963. The feeding habits of Glossina. Bull. World Health Organ. 28, 711–
Wright, S., 1965. The interpretation of population structure by F-statistics with
special regard to system of mating. Evolution 19, 395–420.
J. Bouyer et al./Infection, Genetics and Evolution 10 (2010) 321–328