Genetic diversity and distribution of Senegalia senegal (L.) Britton under climate change scenarios in West Africa
Genetic diversity and distribution of Senegalia senegal (L.) Britton under climate change scenarios in West Africa
Paul Terwase Lyam 0 1 2
JoaquÂõn Duque-Lazo 0 2
Walter Durka 0 2
Frank Hauenschild 1 2
Jan Schnitzler 1 2
Ingo Michalak 1 2
Oluwatoyin Temitayo Ogundipe 2
Alexandra Nora Muellner-Riehl 0 1 2
0 German Centre for Integrative Biodiversity Research (iDiv) Halle-Jena-Leipzig, Germany, 4 Department of Forest Engineering, Laboratory of Dendrochronology , Silviculture and Global Change, DendrodatLab- ERSAF , University of Cordoba, Campus de Rabanales , C oÂrdoba , Spain , 5 Department of Systematic Botany and Functional Biodiversity, Institute of Biology, Leipzig University, Germany, 6 Department of Community Ecology (BZF), Helmholtz Centre for Environmental Research±UFZ , Halle, Germany , 7 Department of Botany, University of Lagos , Akoka, Lagos , Nigeria
1 Department of Molecular Evolution and Plant Systematics & Herbarium (LZ), Institute of Biology, Leipzig University, Germany, 2 National Centre for Genetic Resources and Biotechnology , Ibadan , Nigeria
2 Editor: Tzen-Yuh Chiang, National Cheng Kung University , TAIWAN
Climate change is predicted to impact species' genetic diversity and distribution. We used Senegalia senegal (L.) Britton, an economically important species distributed in the Sudano-Sahelian savannah belt of West Africa, to investigate the impact of climate change on intraspecific genetic diversity and distribution. We used ten nuclear and two plastid microsatellite markers to assess genetic variation, population structure and differentiation across thirteen sites in West Africa. We projected suitable range, and potential impact of climate change on genetic diversity using a maximum entropy approach, under four different climate change scenarios. We found higher genetic and haplotype diversity at both nuclear and plastid markers than previously reported. Genetic differentiation was strong for chloroplast and moderate for the nuclear genome. Both genomes indicated three spatially structured genetic groups. The distribution of Senegalia senegal is strongly correlated with extractable nitrogen, coarse fragments, soil organic carbon stock, precipitation of warmest and coldest quarter and mean temperature of driest quarter. We predicted 40.96 to 6.34 per cent of the current distribution to favourably support the species' ecological requirements under future climate scenarios. Our results suggest that climate change is going to affect the population genetic structure of Senegalia senegal, and that patterns of genetic diversity are going to influence the species' adaptive response to climate change. Our study contributes to the growing evidence predicting the loss of economically relevant plants in West Africa in the
Data Availability Statement: All relevant data are
within the paper and its Supporting Information
Funding: This work was funded by the Alexander
von Humboldt Foundation (AVH), Germany under
the auspices of an International Climate Protection
Fellowship (ICPC), and partly by the German
Academic Exchange Service (DAAD), both granted
to Paul Terwase Lyam and hosted by Alexandra
Muellner-Riehl. Field work and data collection was
next decades due to climate change.
additionally supported by the National Centre for
Genetic resources and Biotechnology (NACGRAB)
Ibadan, Nigeria. Apart from this, the funders have
no role in the study design, subsequent data
collection and analysis, decision to publish or
preparation of the manuscript.
Biodiversity is crucial to human well-being as it provides fundamental ecosystem services. At
the same time, biodiversity is under threat by human population growth, increased land use,
and CO2 emissions, which are all direct or indirect drivers of climate change [1±3]. One of the
regions most threatened by climate change is sub-Saharan Africa [
]. This part of Africa is a
developing region where deforestation and desertification are of major concern. Within
subSaharan Africa lies the so-called arid and semi-arid land traversing the savannah belts of the
region and making up the Sudano-Sahelian zone (SSZ) . This region is socio-economically
and ecologically important, but at the same time threatened by climate change, land
degradation due to unsustainable agriculture, deforestation, and overgrazing [
The effect of environmental and climate change has been studied for a few African plant
3, 8, 9
]. Mostly, such studies focussed on predicting the impact on ecosystems and
]. These studies are highly valuable, but do not account for intra-specific genetic
]. Global studies indicate that large proportions of suitable habitat and the respective
species will be lost during this century due to climate change [
3, 11, 12
]. However, climate
change will also affect intraspecific genetic diversity [
]. Studying the effects of climatic
alterations on genetic diversity is necessary if we are to understand the evolutionary consequences
of climate change and its long-term effects on species distribution [
]. To assess future
impacts on biodiversity many studies have used the Intergovernmental Panel on Climate
Change's (IPCC) scenarios from the Special Report on Emission Scenarios, which project
emissions and socio-economic changes [
]. The latest set of climate model simulations
assumes four different sets of possible futures. They are known as the Representation
Concentration Pathways (RCPs) representing low (RCP 2.6), medium (RCPs 4.5 and 6.0) or high
(RCP 8.5) emissions [
In the sub-Saharan zone, one important species is Senegalia senegal (L.) Britton. (Fabaceae),
syn. Acacia senegal (L.) Willd. [
]. The species, which is known for its exudate named `gum
arabic', is a small deciduous tree, native to arid and semi-desert regions of sub-Saharan Africa,
but has also been introduced in other regions (e.g. the Indian sub-continent) [
western Africa, S. senegal naturally occurs either as a dominant extensive pure stand or in
co-habitation with other species in a variety of vegetation types, including semi-desert grassland,
Anogeissus woodland and rocky hill slopes . The species can grow on sandy, skeletal and
slightly loamy soils. However, it shows preference for coarse-texture soils such as fossil dunes,
with a soil pH of 5 to 8. Although S. senegal has been observed to grow in areas that receive 100
to 950 mm annual rainfall, it thrives best in areas with 300 to 400 mm annual rainfall. The
species can tolerate five to 11 months of drought and can survive temperatures high as 43ÊC, dry
wind and sandstorms, but is highly sensitive to frost. The altitude ranges from 100 to 1700
metres above sea level within the African ASAL. The species is insect-pollinated, and
predominantly out-crossing [
]. Seeds are dispersed at least partly by animals, especially ungulates
. The astringent, emulsifying, film-forming and encapsulating properties of the gum arabic
represents an important economic resource, which is often used in the food-processing,
pharmaceutical, cosmetic, and lithographic ink industries [24±27]. In addition to playing a
preferential role in the context of land degradation and desertification control, S. senegal has been
reported to be resilient to drought and grazing making it an important species in agroforestry
]. The current conservation status of the species is ªleast concernº (IUCN).
However, field observations show that the species is locally under intense pressure from animal
grazing, pest and human activities (personal observation) thus corroborating the earlier report
of Eisa et al. [
]. In addition, climate change will likely further increase the vulnerability of
the species through disturbances of natural habitat. As a consequence, in situ conservation and
2 / 20
future utility of S. senegal could be jeopardized [
], with obvious repercussions for human
livelihoods. Various aspects including soil physicochemical properties , pollination and
breeding systems [
], ecology [
], phylogeography , and population genetics [
in S. senegal have been previously investigated. In S. senegal, intraspecific genetic variation has
been studied both on regional levels [
] and for range-wide phylogeography [
latter shows a recent range expansion to west Africa with genetic diversity reported to be low in
the region [
]. However, no detailed analyses are available for the SSZ of West Africa where
the available phylogeographic analysis is characterized by sampling gaps.
This study aims to fill existing sampling gaps and identify genetic diversity and population
structure of Senegalia senegal (L.) Britton. in the SSZ of West Africa. We used both nuclear and
maternally inherited plastid microsatellite markers. Furthermore, we used species distribution
modelling to forecast the potential range loss and loss of intra-specific genetic diversity under
climate change and promote strategies towards effective management and conservation of S.
senegal in the West African SSZ. Specifically, we address the following questions: 1) Is the
genetic diversity in the West African SSZ as low as previously stated for the region? 2) How is
genetic diversity structured within the West African SSZ? 3) How would future climate impact
the distribution of S. senegal in the West African SSZ? We discuss the potential impact of
climate driven range changes of S. senegal in the SSZ of West Africa on extant genetic variation.
Materials and methods
No specific permissions were required for sampling of leaf material. There is no previous
report stating that Senegalia senegal is threatened or under protection in any of the sampled
locations. According to IUCN, the species has ªleast concernº conservation status.
Study area and sampling
Fresh leaf samples of Senegalia senegal were collected from thirteen localities along the fringes
of the gum arabic belt, situated between latitude 11ÊN and 13ÊN and longitude -0.1ÊW and
13.2ÊE between 2012 and 2016. The study area is a mixture of tropical woodland, grassland
savannah and semi-desert steppe that transverses the SSZ and extends eastwards from
northeastern Ghana to north-eastern Nigeria (Fig 1). The spatial distances between neighbouring
collection sites ranged between 35 km (BIR±GUR) and 1,500 km (BKG±MDG). The sample
size per site was between 13 and 32 individuals, with a total of 316 samples. Within sites
distance was between 10 m and 10 km, depending on the size of the population. Details of
parameters sampled from the sites are shown in Table 1. Field-collected material was dried in silica
gel prior to DNA extraction [
]. At least one individual per population was deposited as a
voucher specimen at both, the herbaria of the National Centre for Genetic Resources and
Biotechnology (NACGRAB), Ibadan, Nigeria, and Leipzig University (LZ), Germany.
Total genomic DNA was extracted from 30 mg of silica-dried leaves using a Nucleo-spin plant
II kit (Macherey-Nagel, DuÈren, Germany). We used ten nuclear microsatellites (hereafter
nSSR) specifically designed for Senegalia senegal by Assoumane et al. [
], and two universal
plastid microsatellites (hereafter cpSSR) designed by Weising et al. : ccmp5 and ccmp10.
DNA amplification was performed using a multiplex PCR kit (Qiagen, Venlo, The
Netherlands). The PCR mix contained a final volume of 10 μl with 5 μl of 2X Qiagen Multiplex PCR
Master Mix, 1 μl of Primer mix (0.2 μM of each forward and reverse primers), 1 μl of
3 / 20
Fig 1. Location of thirteen Senegalia senegal populations sampled from West Africa.
Q-Solution, 1 μl of RNase free water and 2 μl of template DNA (10 ng/μl). The PCR was
initially heated to 95ÊC for 15 min, followed by a 3-step touchdown cycling programme
consisting of denaturation at 95 oC for 15 mins, annealing at 67 oC for 1.5 mins, extension at 72 oC
for 1 min, followed by eight cycles of 94 oC for 30 s, 65 oC for 1.5 min with 2 oC decrease at
each cycle, 72 oC for 1 min; 24 cycles at 94 oC for 30 s, 51 oC for 1.5 min, 72 oC for 1 min, and a
Abbreviations of populations are listed in the first column; number of samples per population (N). Mean annual rainfall and temperature sourced from Worldclim 1.4
and 2.0 respectively.
4 / 20
final extension at 60 oC for 30 min. Amplification was carried out separately for each primer
pair in a singleplex-PCR and combined for fragment analysis. PCR products were amplified
using 6-FAM-, VIC-, NED-, and PET-labelled primers (Applied Biosystems, Foster City,
California, USA). The touchdown cycling program was used on a Mastercycler 2050 model
thermal cycler (Thermo Fisher Scientific, Waltham, Massachusetts, USA). PCR products were
sequenced on an ABI 3730 automated sequencer (Applied Biosystems) using LIZ 500 or LIZ
600. Allelic binning and scoring of genotypes were performed using GeneMapper v.5.0
(Applied Biosystems). Raw data matrices containing allelic information were double-checked
for scoring errors. Samples with more than two missing loci were excluded from further
analysis. The final data set consisted of 287 samples for nSSR and 303 samples for cpSSR analyses.
Genetic data analysis
We used the program MicroChecker v.2.2.3 [
] to estimate the frequency of Null alleles
across all samples. For the nSSR data set, intrapopulation genetic diversity was estimated as the
total number of alleles per locus and per population (NA), the average number of alleles per
locus for each population over all loci (Aavr), the unbiased estimate of expected (HE) and
observed (HO) heterozygosity [
] within populations using Arlequin v.188.8.131.52 [
number of private alleles (APRIV) was estimated using GDA v.1.0 [
]. FSTAT v.1.2 [
] was used
to calculate allelic richness (AR), thus correcting for different sample sizes, and inbreeding
coefficient (FIS). The package genepop v.1.2 [
] was used to perform exact tests of Hardy±
Weinberg equilibrium (HWE), and F-statistics estimated by Weir and Cockerhan .
Population genetic structure was evaluated using a Bayesian clustering approach, implemented in
STRUCTURE v.2.3.4 [
]. Individual's genotype association from 1 to 13 distinct genetic
clusters (K), allowing for admixture, was performed 20 times for each K with 500,000 generations
as burn-in period, followed by 500,000 iterations of MCMC analysis assuming that the
different populations had correlated allele frequencies without using prior information on the
sampling locations of each individual [
]. The ªEvanno methodº [
] was used to estimate the
most likely K [
], as implemented in STRUCTURE HARVESTER [
], but we also
scrutinized results of other K for biologically meaningful solutions. To estimate population genetic
differentiation, an analysis of molecular variance (AMOVA) was performed based on the
clusters suggested by the STRUCTURE analysis using GenAlex v.6. [
]. We tested for isolation by
distance (IBD) based on the correlation between genetic and geographical distances between
populations using a Mantel test as implemented in GenAlex v.6. [
For the cpSSR data, HAPLOTYPE v.1.05 [
] was used to estimate the number of alleles at
chloroplast SSR loci (NacpSSR), number of haplotypes detected in each population (Nb),
effective number of haplotypes (Ne), genetic diversity (Dv), haplotype richness (Hr) and the
number of private haplotypes (Prv). A parsimony network illustrating genetic relationships
between haplotypes was inferred using the software PopArt [
], assuming single-step
mutations between alleles. Total genetic variation among samples was calculated using the
phi-statistics of AMOVA which was performed using GenAlex [
]. This total variation was
partitioned at three levelsÐwithin populations (Phi-PT), among populations within clusters
(Phi-PR), and among clusters (Phi-RT) [
The sampled point locations (n = 316) were initially increased with 30 records from the GBIF
]. The dataset was extensively cleaned with doubtful localities and multiple records
within a 30 arc-second gridcell (approximately 1x1 km, see below) being removed. The final
set of point locations (n = 251) was then used for estimating the potential distribution of the
5 / 20
species. Additionally, pseudo-absence data were created with the same number of points as the
presence data (n = 251). Pseudo-absences were randomly generated across the study area with
a minimum distance of 1.5 km to each other and the presence localities. Equal numbers of
presence and pseudo-absence were used as recommended by Liu et al. [
bioclimatic variables, together with three topographic (DEM; DEM-derived) variables with a spatial
resolution of 30 arc-seconds were downloaded from the public WorldClim database [
These variables describe interpolated monthly means of climatic/environmental factors.
Additional 66 soil variables with a resolution of 250 meters (S5 Table) were downloaded from the
online database Soildgrids250m [
]. The complete set of variables was resampled to 30
arcseconds spatial resolution in the WGS 84 coordinate system.
The impact of climate change on the current distribution of Senegalia senegal was assessed
using WorldClim's future climate change scenarios. We selected two future climate change
projections (2050 and 2070) from two global climate models (GISS-ER2 [
] and CCSM4
, hereafter GCM) and four RCP scenarios [
]. Moreover, topographic and soil variables
were considered constant in future scenarios. Distribution modeling (current and future) of S.
senegal was computed with a maximum entropy approach [
] as implemented in MaxEnt
]. Model performance was evaluated by the Area Under the Curve (AUC) of the
Receiver Operating Characteristic (ROC) plot. The final model was selected as the one with
highest AUC score [
]. Variable selection was performed in two steps: First, collinear
variables were deleted stepwise considering a variance inflation factor (VIF) > 10 as critical
threshold  using the 'usdm' R package [
], reducing the number of variables to n = 12. Second,
the final set of environmental variables was chosen based on the optimization of the AUC
value by the Random Forest (RF) model using the 'AUC-RF' R package [
], resulting in seven
environmental predictors for the final model. In order to assess area gains and losses, we
reclassified current and future habitat suitability probability scores into binary data by a
threshold equal to prevalence [
]. We then estimated the degree of total area change as the
percentage of remaining area in comparison with the current distribution. Values above 100%
represent expansion (gain of area distribution) and values below 100% represent the
diminution of the total extend from the current prediction. In addition, we mapped the changes
between the current and future habitat suitability scores of S. senegal for each 30 arc-second
grid cell. Differences were classified as follows: 1) not present, low probability of occurrence in
current and future predictions (<75%); 2) area loss, high probability in the current prediction
and low in the future; 3) area gain, low probability of occurrence (<75%) in the current
situation and high in future predictions (>75%) and 4) remain, high probability of occurrence
(>75%) as previously done by [
Within-Population genetic diversity
In total, 110 alleles were scored across ten nSSR loci for the 287 genotyped individuals (S1
Table). None of the loci exhibited null allele presence. At the population level, observed and
expected heterozygosity values for nSSR markers ranged from 0.42 to 0.73 (mean 0.61) and
0.43 to 0.63 (mean 0.56), respectively. Allelic richness ranged from 2.86 to 4.55 with a mean of
3.71. Private alleles were identified in eleven populations, with GOU having the highest
number of four private alleles. Significant deviations from HWE were detected at seven sites with
heterozygote deficiency in one site (HAD) and heterozygote excess in six sites (Table 2) while
measures of inbreeding coefficient (Fis) were generally negative or very low ranging from
-0.28 to 0.12 (mean = -0.06).
6 / 20
Number of samples per location (N), allelic richness (AR), mean number of alleles per locus per population (NA), observed heterozygosity (HO), expected heterozygosity
(HE), number of private alleles (APRIV), Fixation index (FIS), number of alleles at chloroplast SSR loci (NacpSSR), number of haplotypes detected in each population (Nb),
effective number of haplotypes (Ne), number of private haplotypes (Prv), haplotypic richness (Hrs), genetic diversity, (DV)
1 p<0.05, p<0.01, p<0.001.
+Mean average across all loci with standard deviation.
Genetic diversity parameters and locus level estimates for cpSSR loci are summarised in
Tables 2 and S4 respectively. Both cpSSR loci were polymorphic, exhibiting three (ccmp5) and
five (ccmp10) alleles per locus among the 303 samples. The combination of alleles resulted in 6
unique haplotypes. Haplotypes H4, H5 and H2 were most frequent occuring in 119, 85 and 79
individuals respectively, whereas H6 was found 14 times and haplotypes H1 and H3 only once.
Fig 2. Bar plots of proportional group membership for the 287 individuals genotyped at 10 nSSR loci. (A). K = 2
and (B). K = 3. Vertical bars represent samples. Lines separate populations with colours representing the proportion of
ancestry derived from each group. Cluster 1 is shown in green, cluster 2 in brown and cluster 3 in orange.
7 / 20
Three haplotypes (H1, H3 and H6) were present in only one population each. Populations
YUS harboured two private haplotypes while BRN harboured one private haplotype (Tables 2
and S3). Between one and four haplotypes were found per population, resulting in haplotype
richness (HR) between 0 and 3.25 (Table 2).
Bayesian analysis of population structure using nSSR suggested the presence of two
genetically distinct clusters based on the Evanno method [
], with individuals showing only
little mixed ancestry (Fig 2A, S1 Fig). However, K = 3 (Fig 2B) cannot be neglected as it had
higher mean likelihood L(K) and still a large value of ΔK. Samples from four geographical
populations (BKG, ZUR, SOK and MDG) were affiliated to cluster 1 (green, cluster
membership coefficient > 96%). This cluster comprised three populations from the
westernmost and one from the eastern-most part of the study area. Samples from nine
geographical populations made up the red cluster at K = 2 (membership coefficient between 65.4
and 98%), which was again separated at K = 3 into two: cluster 2 (MAD, AGU, GOU±
brown) and cluster 3 (RUM, HAD, BRN, GUR, JAK±orange). Population YUS was highly
mixed and shared ancestry between clusters 2 and 3 (49% vs 47%). This separation
reflected a latitudinal pattern with the northern-most populations making up cluster 2
leaving six populations in the central part of the study area as cluster 3. Population YUS
showed mixed ancestry with 49% and 47% affiliation to clusters 2 and 3, respectively (S2
With respect to cpSSR, 12 out of the 13 populations were either fixed to a single cpDNA
haplotype or had a highly dominant haplotype (Fig 3, S3 Table). The population structure at
cpSSR loci closely resembled the one at nSSR. While H2 was present in all four populations
making up cluster 1, H5 was dominant in populations of cluster 2 and H4 in populations that
corresponded to cluster 3. Similar to the nSSR data set, population YUS was mixed of two
haplotypes (H4, H6).
With the nuclear data set, non-hierarchical AMOVA revealed 13% variation among and
87% of genetic variation residing within populations (FST = 0.14; P 0.001). A hierarchical
AMOVA revealed that 18% of the variation resided among clusters at K = 2, while 5% and
77% resided among and within populations (FST = 0.23; P 0.001). Considering three clusters,
differences among the clusters accounted for 13.5% while variation among and within
populations accounted for 4% and 82.5% (FST = 0.18; P 0.001) respectively. By contrast, cpSSR
non-hierarchical data showed 85% variation among and 15% of genetic variation residing
within populations (PhiPT = 0.86; P 0.001). A hierarchical AMOVA (K = 2) found 51% of
the total genetic variance between populations from different clusters (PhiPT = 0.86). Among
populations within clusters contributed 39% of the total genetic variance (PhiPR), and 10% of
the genetic variation was obtained from within individual populations (PhiPT = 0.89). All
three levels contributed significantly to the overall genetic variation as determined through the
permutation analyses (Table 3).
Pairwise FST values between populations from different clusters (as suggested by
STRUCTURE) were generally higher than those between populations from the same cluster
(S6 Table). The Mantel test with 999 permutations revealed that genetic divergence of
populations was correlated with geographic distance (r = 0.53; P = 0.001) indicating that the
isolation by distance model cannot be rejected (Fig 4). Despite this significant IBD pattern, the
geographically very distant but genetically close population MDG within the green nSSR
cluster indicates the presence of a larger-scale phylogeographical structure that is not
consistent with IBD.
8 / 20
Fig 3. Distribution range, sampling sites and genetic structure for Senegalia senegal populations analysed in the present
study. (A) nuclear (B) study area and (C) chloroplast genomes. Each population is represented by a pie chart showing
proportional membership of clusters or share of haplotypes. Haplotype network was generated by TCS in PopArt with circle sizes
proportional to the relative frequency of a particular haplotype.
9 / 20
Percentage molecular variance (% Mol. var.), differentiation among individuals (FST), differentiation among individuals within populations (FIS), differentiation among
populations (FIT) is given.
Support is illustrated by a star ( ) if p 0.001.
PhiRT, proportion of the total genetic variance that is due to the variance between clusters; PhiPR, proportion of the total genetic variance that is due to the variance
among populations within a cluster; PhiPT, proportion of the total genetic variance that is due to the variance among individuals within a variant.
The final response variable data set to model the current distribution of Senegalia senegal
included four climatic and three soil variables (Table 4; S5 Table). The MaxEnt model
predicting the current distribution (Fig 5A and 5B) was statistically supported with an AUC value of
0.958 (S2 Fig). The distribution of S. senegal in West Africa is most strongly correlated with
Fig 4. Relationship between genetic and geographic distances (isolation by distance) of Senegalia senegal in West
10 / 20
Probability of selection
extractable nitrogen, coarse fragments, soil organic carbon stock, precipitation of warmest and
coldest quarter, precipitation of wettest month and mean temperature of driest quarter
(Table 4). Trends in future distribution were similarly forecasted by both GCM but only the
resulting maps from GISSER2 are displayed. Our dataset shows that 35.34% to 11.39% of the
Fig 5. Potential current and future distribution maps for Senegalia senegal across the study area. Location of sample points (A, red triangles) and habitat suitability
map for Senegalia senegal based on present-day climatic conditions (B, brown shaded areas). Predicted potential distribution maps under future conditions: 2050 (C±F)
and 2070 (G±J) is given according to the representative concentration pathway climate scenarios. Yellow areas indicate unsuitable conditions for S. senegal. The
numbers identifying each of the RCPs (C±J) refer to the magnitude of the energy imbalance measured in watts per square meters in the scenario in the year under
11 / 20
current potential range of the species will favourably support the species' ecological
requirements across all the scenarios by the year 2050 (Fig 5C±5F). Consequently, by the year 2070,
projections show that only 40.96% and 6.34% of the species' current potential distribution
might persist favourably under low emission (RCP 2.6) and high emission (RCP 8.5) scenarios,
respectively (Fig 5G±5J). In addition, both GCMs predict a stabilization of habitat loss between
2050 and 2070 under RCP 2.6, while all other scenarios predict additional habitat loss from
2050 to 2070 (S3 Fig).
Within-Population genetic diversity
Genetic diversity is a critical measure in population genetics because it provides insights into
the current and likely future health of a population [
]. Heterozygosity, described as the
average portion of loci with two varying alleles at a single locus within an individual, is a
fundamental measure of genetic diversity [
]. Our analyses revealed high levels of genetic diversity
within populations of Senegalia senegal. Mean expected heterozygosity (HE = 0.56; Table 2)
found in this study is comparable to values found in other microsatellite studies of Senegalia
species (S. senegal: HE = 0.67, [
]; HE = 0.53, [
]; S. dudgeoni, HE = 0.6 [
]; S. tortilis, HE =
]) and other tropical tree species, including Milicia excelsa (HE = 0.81, [
Koompassia malaccensis (HE = 0.85, [
]), and Vitellaria paradoxa (HE = 0.73, [
]). Conversely, low
levels of genetic diversity can lead to inbreeding depression in the short-term, and to reduced
evolutionary potential in the longer term [
Half of the investigated sites significantly deviated from HWE. Moderate FIS values were
found in one population, indicating a heterozygote deficit. Such a reduction can either be
explained by the presence of null alleles [
], the Wahlund effect, or a non-panmictic mating
]. As null alleles or Wahlund effects are not evident in our findings, we attribute the
deviation from HWE at location HAD (FIS = 0.12) to inbreeding and limited gene-flow
consistent with isolation and small population size. On the other hand, heterozygote excess, which
was found in six sites, could potentially be due to selection for heterozygosity (i.e. heterozygote
advantage), or self-incompatibility and obligate outcrossing [
]. This finding coupled with FIS
estimates (mean = 0.06) obtained from our study is in agreement with earlier reports [
suggesting that S. senegal is almost exclusively outcrossed and self-incompatible. Studies on
morphology and genetic variation of some tropical species showed that outcrossed genotypes
also grew faster and had lower mortality than progeny from selfing, resulting in a greater
heterozygosity in the populations due to the selective loss of homozygous individuals [
Furthermore, Wang et al., [
] reported that a species' geographic range and its ecological
attributes influenced genetic diversity; and that high heterozygosity favoured long-lived plants
especially in the arid zones. Our study is congruent with this assumption and is of particular
importance when the historical range shifts Senegalia senegal had experienced [
complex landscapes and fragile ecosystems it occupies [
], and its ecological characteristics, are
taken into consideration [
]. Asynchrony in flowering season has been observed for nearby S.
senegal populations (personal observation), which limits inter-population gene flow [
Perennial insurgence of wild fires, insect pest invasion, soil erosion, overexploitation and
grazing are common ecological challenges throughout these areas [
81, 84, 85
]. These factors
potentially influence or impact gene flow of S. senegal, as shown for several tropical species [
and was further supported by Robledo-Arnuncio et al. . The impact of these factors on
gene flow could have slightly affected the heterozygosity and allele estimates detected in our
study, although S. senegal has been shown to be an outcrossing species [
12 / 20
Population genetic structure and differentiation
There was a strong difference in the extent of genetic structuring revealed by the two types of
markers (FST = 0.14 for nSSRs and FST = 0.86 for cpSSRs), indicating a much stronger
differentiation of the chloroplast genome compared to the nuclear genome. Such differences are
commonly observed and are to be expected due to the different types of inheritance and mode of
]. Low to moderate values of nuclear population differentiation have previously
been documented in Senegalia senegalia and were attributed to likely intense gene flow [
]. However, as FST values only mirror the effects of past events, weak genetic differentiation
among populations could also be observed if populations were young, shared a common
ancestry and were only weakly differentiated despite lack of current gene flow. The high FST
values in the chloroplast data set suggest that gene flow via seeds is negligible, thus highlighting
the importance of pollen vs. seed dispersal for the connectivity of population cohesion in the
In the analysis of STRUCTURE, two clusters (K = 2), representing a southern and a
northern group, may be the most parsimonious explanation for the structure present in the data, yet
we consider it not to offer the most biologically meaningful explanation. This is because three
clusters (K = 3) better reflect the structure in the data, showing groups of populations that are
geographically coherent. Moreover, the three genetic clusters, as revealed by nSSRs were
further fully supported by the analysis of cpDNA (Fig 3). Here, the green cluster matched
haplotype H2, and the brown and orange clusters matched the closely related haplotypes H5 and
H4. Although dispersal via pollen in Senegalia senegal might be effective within its natural
landscapes coupled with the fact that open low density forest facilitates longer pollen dispersal
], human-mediated dispersal through decades of economic usage might have also
influenced the disjunct pattern of populations of cluster 1 observed in the dataset. Because of
its agroforestry importance and high quality gum production, Senegalia seeds have been traded
among nomads, pastorialists and traders of various agricultural products across the region for
]. Various animals such as cattle, sheep, goat, donkey and camels that graze on
the fields for pasture across the entire region play a role in the dispersal of seeds among
population. These human-mediated seed transfer is usually along specific routes of similar arid
environmental condition, such as MDG, ZUR, and BKG. Frequent gene flow is propagated
among those populations without severe interference with populations aside those routes,
such as BRN, JAK, and HAD. In addition, sharp environmental transition due to climate
change and land use has been shown to be responsible for the increasing aridity from north to
south, hence the progressive expansion of Senegalia species from the driest to the wettest zones
in the SSZ [
]. Consequently, this gradient rather than geographic distance accounts for gene
flow among the populations making up cluster 2 and 3 (K = 3), and thus the divergence of
Overall, there is isolation by distance which is in line with gene-flow drift equilibrium and
seed dispersal by animals because with geographical clustering, effects of environment cannot
be totally excluded. Therefore, isolation-by-environment [
] may have as well contributed to
the seemingly disjunct populations. However, because there is congruence in the result
obtained from both nSSR and cpSSR, the whole pattern is very likely to be primarily historic
(seed dispersal), rather than adaptive, but this study cannot establish if the disjunct pattern is
due to isolation by distance or/and isolation by environment.
Odee et al. [
], hypothesized a recent east to west range expansion of African populations
of Senegalia senegal using sequences of ITS and psbA. They illustrated that West African S.
senegal was represented by a single genetic group (nuclear), or two haplotypes (chloroplast),
respectively. However, the data set in the study was characterized by a low resolution for
13 / 20
intraspecific differentiation and contained a sampling gap in the SSZ of West Africa. Our
study unveiled more diversity and population structure in S. senegal (Table 2, Table 3), thus
suggesting a more complex pattern than previously reported for this region.
The detected haplotypes, however, showed relationships amongst them and the populations
to which they were affiliated. Of particular importance is H6 which is quite distinct appearing
in only one population (YUS), with at least six and nine mutations between the two closest
dominant relatives (H2 and H4). Although this study does not account for the origin of
haplotypes, a colonization event probably from MDG to YUS, or somewhere farther from the east,
cannot be neglected.
Genetic diversity and distribution under climate change
The MaxEnt model confirms the role of soil and climate properties in shaping the distribution
of Senegalia senegal in the SSZ, as reported by Traore et al. [
]. Given the future climate
scenarios (Fig 5C±5J), there will be likely a reduction of suitable habitats and population sizes,
and potentially a loss of private alleles and private haplotypes leading to declining genetic
diversity. All four future climate scenarios vary in magnitude of energy imbalance and the
impact of the variation is reflected in the model forecast, as the extent of reduction in suitable
range is largely dependent on the species' ecological requirements. The observed predicted
pattern shows a progressive reduction (RCP2.6 > RCP4.5 > RCP6.0 > RCP8.5) of the current
potential distribution of S. senegal within the study area. This shows that S. senegal may be
under threat due to environmental change in the study area and the species might be facing an
increased risk of local extinction. However, considering the high allele diversity found in this
study, populations may show a predisposition to maintain genetic diversity which may
ultimately be a favorable potential for adaptability and persistence should they disperse to track a
new suitable range. Consequently, the ability of S. senegal to persist in an increasingly marginal
range will depend on factors such as ecological fitness, genetic make-up and dispersal ability.
However, there is no certainty that the dispersal rates of plant species could keep pace with fast
rates of environmental change [
Our study unveiled a high genetic and haplotypic variation in the West African
Sudano-Sahelian range of Senegalia senegal in contrast to previous report that found this part of the range
genetically homogenous. We suggest that the species' colonization history might have played
an important role in the distribution of genetic diversity and structuring of populations in the
study area. Distribution modeling indicates that depending on the climate scenario, only 41%
(RCP 2.6), 35% (RCP 4.5) or down to only 6.3 (RCP 8.5) of suitable area will be available.
Consequently, a large part of the geographically structured genetic variation is threatened. In
particular the genotypes distributed to the West of the study area, microsatellite cluster 1 and
chloroplast haplotype 2, may face strong declines, as hardly any surviving S. senegal are
predicted in this region. Our findings therefore offer insights into how to manage threats or
predict responses to disturbances in relation to environmental changes. Finally, the findings of
this study will provide a valuable base to reinforce the information available to conservationists
and policy makers.
S1 Table. Genetic characteristics of ten nuclear SSR loci in 13 populations of Senegalia senegal (N = 297).
14 / 20
S2 Table. Proportion of membership of each predefined population in each of the inferred clusters at both K = 2 and K = 3.
S3 Table. List of haplotypes detected at two cpSSR loci in 13 populations of Senegalia sene
gal. Population YUS is characterized by 2 two unique haplotypes that are not present in any
other population. Private haplotypes are highlighted in grey/bold.
S4 Table. Genetic characteristics of two chloroplast microsatellite markers and result of
genotyping in S. senegal (N = 303).
S5 Table. List of variables used for modelling the distribution of Senegalia senegal in West
S6 Table. Matrix of pairwise FST values based on allele data at ten nuclear (above diagonal)
and two chloroplast SSR (above diagonal) loci among 13 populations of Senegalia senegal.
Most of the values estimated for both marker types were significant (P 0.001). For
abbreviations of populations see Table 1.
S1 Fig. Plots for detecting the number of K groups that best fit the data according to .
(A) Delta K and (B) Plot of mean likelihood L (K) and variance per K value from
STRUCTURE on a dataset containing 287 individuals genotyped for ten nSSR.
S2 Fig. Plot of sensitivity vs. specificity. Area under the curve (AUC) value of 0.958 indicates
the accuracy of the model prediction.
S3 Fig. Predicted area (%) gain or loss of Senegalia senegal relative to the current distribu
tion under four future climate scenarios (RCP's) as estimated by two models (CCSM4,
GISSER2) for 2050 and 2070.
We are grateful to Stephen Akpalu (CSIR-FORIG, Ghana), Moussa Massaoudou (INRAN,
Maradi-Niger), and Kafil Ibrahim (Rumah-Batsari LGA, Kastina), Olabisi Alamu, Olubuiyi
Mayowa (NACGRAB), Ali Abba-Gana Benisheikh (UNIMAID) and Bulama Gado for
assisting with sample collections in the field. We thank Dr. Peter Otto for his role in facilitating our
sampling efforts, and Dr. Adrien Favre for comments on a previous version of this manuscript.
We appreciate all members of the working group ªMolecular Evolution and Plant Systematics
& Herbarium (LZ)º at Leipzig University, especially Jana Ebersbach and Claudia KruÈger, for
suggestions and support in the molecular laboratory.
Conceptualization: Paul Terwase Lyam, Oluwatoyin Temitayo Ogundipe, Alexandra Nora
15 / 20
Data curation: Paul Terwase Lyam, JoaquÂõn Duque-Lazo, Ingo Michalak.
Formal analysis: Paul Terwase Lyam, JoaquÂõn Duque-Lazo, Walter Durka, Frank
Hauenschild, Ingo Michalak.
Funding acquisition: Paul Terwase Lyam, Alexandra Nora Muellner-Riehl.
Investigation: Paul Terwase Lyam, JoaquÂõn Duque-Lazo, Walter Durka, Frank Hauenschild,
Jan Schnitzler, Ingo Michalak, Oluwatoyin Temitayo Ogundipe, Alexandra Nora
Methodology: Paul Terwase Lyam, JoaquÂõn Duque-Lazo, Walter Durka, Frank Hauenschild.
Project administration: Paul Terwase Lyam, Jan Schnitzler, Alexandra Nora Muellner-Riehl.
Resources: Paul Terwase Lyam, Walter Durka, Oluwatoyin Temitayo Ogundipe, Alexandra
Supervision: Walter Durka, Jan Schnitzler, Ingo Michalak, Oluwatoyin Temitayo Ogundipe,
Alexandra Nora Muellner-Riehl.
Writing ± original draft: Paul Terwase Lyam, JoaquÂõn Duque-Lazo.
Writing ± review & editing: Paul Terwase Lyam, JoaquÂõn Duque-Lazo, Walter Durka, Frank
Hauenschild, Jan Schnitzler, Ingo Michalak, Oluwatoyin Temitayo Ogundipe, Alexandra
16 / 20
17 / 20
Weising K, Gardner RC. A set of conserved PCR primers for the analysis of simple sequence repeat
polymorphisms in chloroplast genomes of dicotyledonous angiosperms. Genome 1999; 42(1):9±19.
18 / 20
19 / 20
1. Sala EO , Chapin FS , Armesto JJ , Berlow E , Bloomfield J , Dirzo R , et al. Global Biodiversity Scenarios for the Year 2100. Science 2000 ; 287 ( 1770 ): 1770 ± 4 .
2. Leadley P , Pereira H , Alkemade R , Fernandez-ManjarreÂs J , ProencËa V , Scharlemann J et al. Biodiversity scenarios: Projections of 21st century change in biodiversity, and associated ecosystem services . MontreÂal; 2010. CBD technical series.
3. Heubes J , Schmidt M , Stuch B , Marquez J , Wittig R , Zizka Gea. The projected impact of climate and land use change on plant diversity: An example from West Africa . J Arid Environ 2013 ; 96 : 48 ± 54 .
4. Pachauri RK , Meyer LA. Climate change: Synthesis report. Contribution of working groups I, II and III to the fifth Assessment report of the Intergovernmental Panel on Climate Change . Geneva; 2015 ;
5. Parry ML , Canziani O , Palutikof J , van der Linden P , Hanson CE . Climate change: Impacts, adaptation and vulnerability. contribution of Working Group II to the fourth assessment report of the Intergovernmental Panel on Climate Change . Cambridge: Cambridge University Press; 2007 . Available from: URL: http://www.loc.gov/catdir/enhancements/fy0827/2008295434-d.html.
6. Bocksberger G , Schnitzler J , Chatelain C , Daget P , Janssen T , Schmidt M , et al. Climate and the distribution of grasses in West Africa . Journal of Vegetation Science 2016 ; 27 ( 2 ): 306 ± 17 .
7. Omondi SF , Kireger E , Dangasuk OG , Chikamai B , Odee DW , Cavers S et al. Genetic Diversity and Population Structure of Acacia senegal (L) Willd . in Kenya. Tropical Plant Biol . 2010 ; 3 ( 1 ): 59 ± 70 .
8. Zerbo I , Bernhardt-RoÈmermann M , OueÂdraogo O , Hahn K , Thiombiano A. Effects of Climate and Land Use on Herbaceous Species Richness and Vegetation Composition in West African Savanna Ecosystems . J Bot 2016 ; 2016(2):1 ± 11 .
9. Serdeczny O , Adams S , Baarsch F , Coumou D , Robinson A , Hare W , et al. Climate change impacts in Sub-Saharan Africa: From physical changes to their social repercussions . Reg Environ Change 2016 ; 17 ( 6 ): 1585 ± 600 .
10. Pfenninger M , Balint M , Pauls SU . Methodological framework for projecting the potential loss of intraspecific genetic diversity due to global climate change . BMC Evol Biol 2012 ; 12 ( 1 ): 224 .
11. McClean CJ , Lovett JC , Kuper W , Hannah L , Sommer JH , Barthlott W et al. African plant diversity and climate change . Ann Mo Bot Gard 2005 ; 92 ( 2 ): 139 ± 52 .
12. Sommer JH , Kreft H , Kier G , Jetz W , Mutke J , Barthlott W. Projected impacts of climate change on regional capacities for global plant species richness . Proc Biol Sci 2010 ; 277 ( 1692 ): 2271 ± 80 . https:// doi.org/10.1098/rspb. 2010 .0120 PMID: 20335215
13. Heywood HV . Global biodiversity assessment . Cambridge: Cambridge University Press Cambridge; 1995 .
14. Pauls SU , Nowak C , Balint M , Pfenninger M. The impact of global climate change on genetic diversity within populations and species . Mol Ecol 2013 ; 22 ( 4 ): 925 ± 46 . https://doi.org/10.1111/mec.12152 PMID: 23279006
15. Nakicenovic N , Davidson O , Davis G , GruÈbler A , Kram T , Lebre La Rovere E. Emissions scenarios: Summary for policymakers; a special report of IPCC Working Group III. Geneva, Switzerland; 2000 .
16. van Vuuren DP , Edmonds J , Kainuma M , Riahi K , Thomson A , Hibbard K et al. The representative concentration pathways: an overview . Climatic Change 2011 ; 109 :5± 31 .
17. Arnell NW , Lloyd-Hughes B . The global-scale impacts of climate change on water resources and flooding under new climate and socio-economic scenarios . Climatic Change 2014 ; 122 : 127 ± 40 .
18. Kyalangalilwa B , Boatwright JS , Daru BH , Maurin O , van der Bank M. Phylogenetic position and revised classification of Acacia s.l. (Fabaceae: Mimosoideae) in Africa, including new combinations in Vachellia and Senegalia . Bot J Linn Soc 2013 ; 172 ( 4 ): 500 ± 23 .
19. Bernhard-Reversat F , Poupon H . Nitrogen cycling in a soil-tree system in a sahelian savanna: example of Acacia senegal . Proceedings of a workshop arranged by the SCOPE/UNEP 1980 : 363 ± 9 .
20. Beshai AA . The economics of a primary commodity: Gum Arabic . Oxf Bull Econ Stat 1984 ; 46 ( 4 ): 371 ± 81 .
21. Orwa C , Mutua A , Kindt R , Jamnadass R , Simons A . Agroforestree database: a tree reference and selection guide version 4 .0. 2009 . Available from: URL: http://www.worldagroforestry.org/af/treedb.
22. Fagg C , Allison GE . Acacia senegal and the gum Arabic trade. Monograph and annotated bibliography . Tropical Forestry Papers No. 42 , Alden Group , Oxford 2004 : 261 .
23. Odee DW , Telford A , Wilson J , Gaye A , Cavers S . Plio-Pleistocene history and phylogeography of Acacia senegal in dry woodlands and savannahs of sub-Saharan tropical Africa: evidence of early colonisation and recent range expansion . Heredity 2012 ; 109 ( 6 ): 372 ± 82 . https://doi.org/10.1038/hdy. 2012 .52 PMID: 22929152
24. Brenan JP . Manual on taxonomy of Acacia species: present taxonomy of four species of Acacia (A. albida, A. senegal, A. nilotica, A . tortilis). Rome: FAO; 1983 .
25. Jøker D. Acacia senegal (L.) Willd. Seed Leaflet-Danida Forest Seed Centre 2000 ; (5).
Maundu P , TengnaÈs B. Useful trees and shrubs for Kenya . ICRAF Technical handbook series 2005 .
27. Raddad EAY , Salih AA , Fadl MAE , Kaarakka V , Luukkanen O . Symbiotic nitrogen fixation in eight Acacia senegal provenances in dryland clays of the Blue Nile Sudan estimated by the 15 N natural abundance method . Plant and Soil 2005 ; 275 ( 1 ): 261 ± 9 .
28. Traore S. , Zerbo L. , Schmidt M. , Thiombiano L . Acacia communities and species responses to soil and climate gradients in the Sudano-Sahelian zone of West Africa . J Arid Environ 2012 ; 87 : 144 ± 52 .
29. Eisa MA , Roth M , Sama G., editor. Acacia senegal (gum Arabic tree) : Present role and need for future conservation/Sudan . Hohenheim: Deutscher Tropentag; 2008 .
30. Githae EW , Gachene CKK , Njoka JT . Soil physicochemical properties under Acacia senegal varieties in the dryland areas of Kenya . Afri J Plant Sci 2011 ; 5 ( 8 ): 475 ± 82 .
31. Mulumba JW , Nyakaana S , Jamnadas R . Genetic Diversity and structure of Acacia senegal (L.) Willd . in Uganda. Int. J. Botany 2012 ; 3 ( 8 ): 107 ± 16 .
32. Tandon R , Shivanna KR , Ram HM . Pollination biology and breeding system of Acacia senegal . Bot J Linn Soc 2001 ; 135 ( 3 ): 251 ± 62 .
33. Gray A , Odee D , Cavers S , Wilson J , Telford A , Grant F et al. Does geographic origin dictate ecological strategies in Acacia senegal (L.) Willd.?: Evidence from carbon and nitrogen stable isotopes . Plant Soil 2013 ; 369 ( 1 ±2): 479 ± 96 .
34. Chiveu C , Dangasuk OG , Omunyin M , Wachira FN . Genetic diversity in Kenyan populations of Acacia senegal (L:) Willd revealed by combined RAPD and ISSR markers . Afr. J. Biotechnol . 2008 ; 7 ( 14 ): 2333 ± 40 .
35. Habeballa RS , Hamza NB , El Gaali EI . Genetic variability in Sudanese Acacia senegal (L.) assessed by random amplified polymorphic DNA . Afr. J. Biotechnol . 2010 ; 9 ( 30 ): 4655 ± 60 .
36. Chase MW , Hills HH . Silica gel: an ideal material for field preservation of leaf samples for DNA studies . Taxon 1991 : 215 ± 20 .
37. Assoumane AA , Vaillant A , Mayaki AZ , Verhaegen D. Isolation and characterization of microsatellite markers for Acacia senegal (L.) Willd., a multipurpose arid and semi-arid tree . Mol Ecol Resour 2009 ; 9 (5):1380±3 . https://doi.org/10.1111/j.1755- 0998 . 2009 . 02669 . x PMID : 21564912
39. van Oosterhout C , Hutchinson WF , Wills DPM , Shipley P. MICRO-CHECKER : Software for identifying and correcting genotyping errors in microsatellite data . Mol Ecol Resour 2004 ; 4 ( 3 ): 535 ± 8 .
40. Nei M. Estimation of average heterozygosity and genetic distance from a small number of individuals . Genetics 1978 ; 89 ( 3 ): 583 ± 90 . PMID: 17248844
41. Excoffier L , Lischer HEL . Arlequin suite ver 3.5: a new series of programs to perform population genetics analyses under Linux and Windows . Mol Ecol Resour 2010 ; 10 ( 3 ): 564 ±7. https://doi.org/10.1111/j. 1755- 0998 . 2010 . 02847 . x PMID : 21565059
42. Lewis , PO , Zaykin D. Genetic data analysis: computer program for the analysis of allelic data 2001 . Available from: URL: http://lewis.eeb.uconn.edu/lewishome/software.html.
43. Goudet J. FSTAT ( Version 1 .2) : AComputer Program to Calculate F-Statistics. J Hered 1995 ; 6 ( 86 ): 245 ± 6 .
44. Raymond M , Rousset F. GENEPOP (version 1.2): population genetics software for exact tests and ecumenicism . J Hered 1995 ; 86 ( 3 ): 248 ± 9 .
Weir BS , Cockerham CC . Estimating F-Statistics for the Analysis of Population Structure . Evolution 1984 ; 38 ( 6 ): 1358 . https://doi.org/10.1111/j.1558- 5646 . 1984 .tb05657. x PMID: 28563791
46. Pritchard J , Stephens M , Donelly P . Inference of Population Structure Using Multilocus Genotype Data . Genetics 2000 ; (155): 945 ± 59 .
47. Falush D , Stephens M , Pritchard JK . Inference of population structure using multilocus genotype data: linked loci and correlated allele frequencies . Genetics 2003 ; 164 ( 4 ): 1567 ± 87 . PMID: 12930761
48. Evanno G , Rednaut S , Goudet J . Detecting the number of clusters of individuals using the software structure: A simulation study . Mol Ecol 2005 ; 14 ( 8 ): 2611 ± 20 . https://doi.org/10.1111/j. 1365 - 294X . 2005 . 02553 . x PMID : 15969739
49. Earl DA , vonHoldt BM . STRUCTURE HARVESTER: A website and program for visualizing STRUCTURE output and implementing the Evanno method . Conserv Genet Resour 2012 ; 4 ( 2 ): 359 ± 61 .
50. Peakall ROD SP. GENALEX 6: genetic analysis in Excel. Population genetic software for teaching and research . Mol Ecol Notes 2006 ; 6 ( 1 ): 288 ± 95 .
51. Eliades NG , Eliades DG . Haplotype Analysis: Software for analysis of haplotype data . Forest Genetics and Forest Tree Breeding , Georg-August University Goettingen, Germany. 2009 .
52. Leigh JW BD. PopArt: Full-feature software for haplotype network construction . Methods Ecol Evol 2015 ; 6 ( 9 ): 1110 ± 6 .
53. Xu J , Sha T , Li YC , Zhao ZW , Yang ZL . Recombination and genetic differentiation among natural populations of the ectomycorrhizal mushroom Tricholoma matsutake from southwestern China . Mol Ecol 2008 ; 17 ( 5 ): 1238 ± 47 . https://doi.org/10.1111/j. 1365 - 294X . 2007 . 03665 . x PMID : 18302686
54. GBIF.org. GBIF Occurrence Download http://doi.org/10.15468/dl.c8atij 9th November 2016 .
55. LIU C , Berry PM , Dawson TP , Pearson RG . Selecting thresholds of occurrence in the prediction of species distributions . Ecography 2005 ; 28 ( 3 ): 385 ± 93 .
56. Hijmans RJ , Cameron SE , Parra JL , Jones PG , Jarvis A. Very high resolution interpolated climate surfaces for global land areas . Int. J. Climatol . 2005 ; 25 ( 15 ): 1965 ± 78 .
57. Hengl T , Mendes de Jesus J , Heuvelink GBM , Ruiperez Gonzalez M , Kilibarda M , Blagotic A et al. SoilGrids250m: Global gridded soil information based on machine learning . PLoS ONE 2017 ; 12 ( 2 ):1± 40 .
58. Miller RL , Schmidt GA , Nazarenko LS , Tausnev N , Bauer SE , DelGenio AD et al. CMIP5 historical simulations (1850±2012) with GISS ModelE2 . J Adv Model Earth Syst 2014 ; 6 ( 2 ): 441 ± 78 .
59. Nazarenko L , Schmidt GA , Miller RL , Tausnev N , Kelley M , Ruedy R et al. Future climate change under RCP emission scenarios with GISS ModelE2 . J Adv Model Earth Syst 2015 ; 7 ( 1 ): 244 ± 67 .
60. Gent PR , Danabasoglu G , Donner LJ , Holland MM , Hunke EC , Jayne SR et al. The community climate system model version 4 . J Clim 2011 ; 24 ( 19 ): 4973 ± 91 .
61. Phillips SJ , Anderson RP , Schapire RE . Maximum entropy modeling of species geographic distributions . Ecol Modell 2006 ; 190 ( 3 ±4): 231 ± 59 .
62. Phillips SJ , DudÂõk M. Modeling of species distributions with Maxent: New extensions and a comprehensive evaluation . Ecography 2008 ; 31 ( 2 ): 161 ± 75 .
63. Fielding AH , Bell JF . A review of methods for the assessment of prediction errors in conservation presence/absence models . Environ Conserv 1997 ; 24 ( 1 ): 38 ± 49 .
64. Duque-Lazo J , van Gils H , Groen TA , Navarro-Cerrillo RM . Transferability of species distribution models: The case of Phytophthora cinnamomi in Southwest Spain and Southwest Australia . Ecol Modell 2016 ; 320 : 62 ± 70 .
65. Quinn GP , Keough MJ . Experimental design and data analysis for biologists . Cambridge: Cambridge University Press; 2002 .
66. Naimi B. usdm: Uncertainty analysis for species distribution models . R package version 2013 ; 1 :1± 12 .
67. Calle ML , Urrea V Boulesteix A-L , Malats N. AUC-RF: a new strategy for genomic profiling with random forest . Hum Hered 2011 ; 72 ( 2 ): 121 ± 32 . https://doi.org/10.1159/000330778 PMID: 21996641
68. Duque-Lazo J , Navarro-Cerrillo RM . What to save, the host or the pest? The spatial distribution of xylophage insects within the Mediterranean oak woodlands of Southwestern Spain . For Ecol Manage 2017 ; 392 : 90 ± 104 .
69. Navarro-Cerrillo RM , RumbaoÂ IC , Vidaña AL , PeÂrez J , Duque-Lazo J. IntegracioÂn de datos de inventario y modelos dehaÂbitat para predecir la regeneracioÂn de especies leñosas mediterraÂneas en repoblaciones forestales . Revista Ecosistemas 2016 ; 25 ( 3 ):6± 21 .
70. Kirk H , Freeland JR . Applications and implications of neutral versus non-neutral markers in molecular ecology . Int J Mol Sci 2011 ; 12 ( 6 ): 3966 ± 88 . https://doi.org/10.3390/ijms12063966 PMID: 21747718
71. Nei M. Analysis of gene diversity in subdivided populations . Proc Natl Acad Sci USA 1973 ; 70 ( 12 ): 3321 ± 3 . PMID: 4519626
72. Assoumane A , Mayaki AZ , Favreau B , Bezancon G , Saadou M , Verhaegen D . Differentiation between two sub-species of Acacia senegal complex: Acacia senegal (L.) Willd. and Acacia dudgeoni Craib ex Holland using morphological traits and molecular markers . Genet Resour Crop Evol 2012 ; 59 ( 1 ): 39 ± 48 .
73. Winters G , Shklar G , Korol L . Characterizations of microsatellite DNA markers for Acacia tortilis . Conserv Genet Resour 2013 ; 5 ( 3 ): 807 ± 9 .
74. Ouinsavi C , Sokpon N , Khasa DP . Genetic Diversity and Population Structure of a Threatened African Tree Species, Milicia excelsa, Using Nuclear Microsatellites DNA Markers . International Journal of Forestry Research 2009 ; 2009 (1345): 1 ± 8 .
75. Noreen AME , Webb EL . High Genetic Diversity in a Potentially Vulnerable Tropical Tree Species Despite Extreme Habitat Loss . PLoS ONE 2013 ; 8 ( 12 ).
76. Gwali S , Vaillant A , Nakabonge G , Okullo JL , Eilu G , Muchugi A et al. Genetic diversity in shea tree (Vitellaria paradoxa subspecies nilotica) ethno-varieties in Uganda assessed with microsatellite markers . Forests, Trees and Livelihoods 2014 ; 24 ( 3 ): 163 ± 75 .
77. Poczai P , Varga I , Bell NE , Hyvonen J . Genomics Meets Biodiversity: Advances in Molecular Marker Development and their Applications in Plant Genetic Diversity Assessment . In: Caliskan M, editor. The Molecular Basis of Plant Genetic Diversity: INTECH Open Access Publisher; 2012 . pp 3± 31 .
78. Stoeckel S , Grange J , Fernandez-Manjarres JF , Bilger I , Frascaria-Lacoste N , Mariette S. Heterozygote excess in a self-incompatible and partially clonal forest tree speciesÐPrunus avium L . Mol Ecol 2006 ; 15 ( 8 ): 2109 ± 18 . https://doi.org/10.1111/j. 1365 - 294X . 2006 . 02926 . x PMID : 16780428
79. Liu W , Zhao Y , You J , Qi D , Zhou Y , Chen J et al. Morphological and Genetic Variation along a North-toSouth Transect in Stipa purpurea, a Dominant Grass on the Qinghai-Tibetan Plateau: Implications for Response to Climate Change . PLoS ONE 2016 ; 11 ( 8 ):1± 22 .
80. Wang Z , Kang M , Liu H , Gao J , Zhang Z , Li Y et al. High-level genetic diversity and complex population structure of Siberian apricot (Prunus sibirica L.) in China as revealed by nuclear SSR markers . PLoS ONE 2014 ; 9 ( 2 ):1± 13 .
81. Chidumayo EN , Gumbo DJ . The dry forests and woodlands of Africa: Managing for products and services . London, UK: Earthscan; 2010 .
82. Bouchenak-Khelladi Y , Maurin O , Hurter J , van der Bank M. The evolutionary history and biogeography of Mimosoideae (Leguminosae): An emphasis on African acacias . Mol Phylogenet Evol 2010 ; 57 ( 2 ): 495 ± 508 . https://doi.org/10.1016/j.ympev. 2010 . 07 .019 PMID: 20696261
83. Rozzi R , Arroyo MK , Amesto JJ . Ecological factors affecting gene flow between populations of Anarthrophyllum cumingii (Papilionaceae) growing on equatorial- and polar-facing slopes in the Andes of Central Chile . Plant Ecol 1997 ; 132 : 171 ± 9 .
84. Mujawamariya G , Madi OP , Zoubeirou AM , Sene A , Maisharou A , D'Haese M. Common challenges in gum arabic production and commercialization in West Africa: a comparative study of Cameroon, Niger and Senegal . International Forestry Review 2013 ; 15 ( 2 ): 182 ± 99 .
85. Gaisberger H , Kindt R , Loo J , Schmidt M , Bognounou F , Da SS , Diallo OB , et al. Spatially explicit multithreat assessment of food tree species in Burkina Faso: A fine-scale approach . PLoS ONE 2017 ; 12 ( 9 ):1± 26 .
86. Hamilton B . Tropical tree gene flow and tree dispersal: deforestation affects the genetic structure of the surviving forest fragments . Nature 1999 ; 401 : 129 ± 30 .
87. Robledo-Arnuncio JJ , Klein EK , Muller-Landau HC , SantamarÂõa L. Space, time and complexity in plant dispersal ecology . Mov Ecol 2014 ; 2 ( 1 ):1± 17 . https://doi.org/10.1186/2051-3933-2-1
88. Petit RJ , Duminil J , Fineschi S , Hampe A , Salvini D , Vendramin GG . Comparative organization of chloroplast, mitochondrial and nuclear diversity in plant populations . Mol Ecol 2005 ; 14 ( 3 ): 689 ± 701 . https:// doi.org/10.1111/j. 1365 - 294X . 2004 . 02410 . x PMID : 15723661
89. GoncËalves-Oliveira RC , WoÈhrmann T , Benko-Iseppon AM , Krapp F , Alves M , Wanderley ML et al. Population genetic structure of the rock outcrop species Encholirium spectabile (Bromeliaceae): The role of pollination vs. seed dispersal and evolutionary implications . Am J Bot 2017 ; 104 ( 6 ): 868 ± 78 . https://doi. org/10.3732/ajb.1600410 PMID: 28611073
90. Sexton JP , Hangartner SB , Hoffmann AA . Genetic isolation by environment or distance: Which pattern of gene flow is most common? Evolution 2014 ; 68 ( 1 ):1± 15 . https://doi.org/10.1111/evo.12258 PMID: 24111567
91. Malcolm JR , Markham A , Neilson RP , Garaci M. Estimated migration rates under scenarios of global climate change . J Biogeogr 2002 ; 29 ( 7 ): 835 ± 49 .