Spatial genetic structure within populations and management implications of the South American species Acacia aroma (Fabaceae)
Spatial genetic structure within populations and management implications of the South American species Acacia aroma (Fabaceae)
Carolina Pometti 0 1
Cecilia Bessega 0 1
Ana Cialdella 1
Mauricio Ewens 1
Beatriz Saidman 0 1
Juan Vilardi 0 1
0 Universidad de Buenos Aires, Facultad de Ciencias Exactas y Naturales, Departamento EcologÂõa, Gen eÂtica y EvolucioÂn, Gen eÂtica de Especies Leñosas (GEEL) , Buenos Aires , Argentina , 2 CONICET- Universidad de Buenos Aires, Instituto de EcologÂõa, Gen eÂtica y EvolucioÂ n (IEGEBA) , Buenos Aires, Argentina, 3 IBODA, CONICET, San Isidro, Buenos Aires , Argentina , 4 Estaci oÂn Experimental Fern aÂndez- UCSE (Convenio Provincia Sgo del Estero- Universidad Cat oÂlica Sgo del Est.), Departamento de Robles , Santiago del Estero , Argentina
1 Editor: Giovanni G Vendramin, Consiglio Nazionale delle Ricerche , ITALY
The identification of factors that structure intraspecific diversity is of particular interest for biological conservation and restoration ecology. All rangelands in Argentina are currently experiencing some form of deterioration or desertification. Acacia aroma is a multipurpose species widely distributed throughout this country. In this study, we used the AFLP technique to study genetic diversity, population genetic structure, and fine-scale spatial genetic structure in 170 individuals belonging to 6 natural Argentinean populations. With 401 loci, the mean heterozygosity (HE = 0.2) and the mean percentage of polymorphic loci (PPL = 62.1%) coefficients indicated that the genetic variation is relatively high in A. aroma. The analysis with STRUCTURE showed that the number of clusters (K) was 3. With Geneland analysis, the number of clusters was K = 4, sharing the same grouping as STRUCTURE but dividing one population into two groups. When studying SGS, significant structure was detected in 3 of 6 populations. The neighbourhood size in these populations ranged from 15.2 to 64.3 individuals. The estimated gene dispersal distance depended on the effective population density and disturbance level and ranged from 45 to 864 m. The combined results suggest that a sampling strategy, which aims to maintain a considerable part of the variability contained in natural populations sampled here, would include at least 3 units defined by the clusters analyses that exhibit particular genetic properties. Moreover, the current SGS analysis suggests that within the wider management units/provinces, seed collection from A. aroma should target trees separated by a minimum distance of 50 m but preferably 150 m to reduce genetic relatedness among seeds from different trees.
Data Availability Statement: All relevant data are
within the paper and its Supporting Information
frontend/agencia/fondo/foncyt). The funders had
no role in study design, data collection and
analysis, decision to publish, or preparation of the
The spatial structure of the genomic variation among natural populations constitutes a central
topic in evolutionary biology. The structure is primarily influenced by the population density,
breeding system, and environmental heterogeneity, among other factors. For plants, the ability
to extend the geographical distribution and maintain genetic variability within populations
depends on the gene flow mediated by seed movement and pollen dispersal [
mechanisms influence the structuration of genetic diversity within and between populations, which
is usually referred to as spatial genetic structure (SGS) . The spatial distribution of
individuals within a population is a considerable determinant of population genetic structure, and this
is affected by dispersal processes. Consequently, studying the causes of SGS in plant species, in
particular in Acacia aroma due to its ecological and economic importance in South America,
is useful in conservation and management strategies for maintaining genetic diversity,
particularly in stages of rapid habitat degradation [
Habitat fragmentation and degradation may reduce the size and increase the spatial
isolation of plant populations. This could lead to increased random genetic drift, elevated
inbreeding and reduced gene flow [
4, 5, 6
]. Therefore, when habitat degradation is occurring, it is
important to measure baseline SGS to inform management decisions aimed at maintaining
genetic variation [
Acacia s.l. is the second largest genus in the family Leguminosae (Fabaceae) and currently
comprises more than 1,450 species in three subgenera. Additionally, in many dryland areas,
Acacia s.l. is the dominant shrub or tree on which humans and animals depend, for example A.
nilotica in Africa, A. farnesiana in Mexico, and A. aroma in Argentina, Bolivia and Peru [
Acacia aroma Gillies ex Hook. & Arn. is an ecologically and economically valuable species
facing considerable habitat degradation and is well distributed throughout the Chaco Region,
where it is commonly known as ªaromitoº or ªtuscaº. This species is an outcrosser; it is
pollinated by bees [
], and its seeds are dispersed by native mammals, livestock, goats and horses
]. The deterioration of landscapes in Argentina began two centuries ago, and although the
human population was very small, overgrazing without regard for environmental impact was
standard practice. The ecosystems were fragile and prone to extensive damage, and
consequently, Argentine rangelands underwent desertification. The Chaco Region in Northwest
Argentina experienced severe forest exploitation and overgrazing. Cattle production
progressively expanded in this area, and due to overgrazing, resulted in a shift from grasses to shrubs
and saplings according to the availability. Later, goats replaced cattle as they were able to eat
almost any plant, thereby removing the community from the climax stage. Currently in this
region, some Acacia s.l. species occur together with other legumes, and they provide pods for
forage, lumber and medicinal products [
A. aroma grows as a tree or shrub 4±6 m in height, which is economically and ecologically
important because of multiple uses extending from its roots to its pollen. From an ecological
point of view, the roots are good nitrogen fixers, and the fruits and leaves provide forage for
cattle and goats. Economically, the fruits and bark are rich in tannins, the flowers are useful in
the perfume industry, the seeds have medicinal uses and the pollen is used in honey
]. Currently, there are few studies on population structure and genetic diversity in
natural populations of A. aroma.
Thus, our main objectives are to analyse the genetic diversity and structure and characterize
fine-scale spatial genetic structure (SGS) in five natural populations and a remnant of A.
aroma in the Chaco Biogeographical Region in Northwest Argentina, each with different levels
of disturbance. The results of this study can be used to facilitate the sustainable management
of A. aroma, and this methodology can be extended to other Acacia species.
2 / 14
Materials and methods
Young, fully grown, healthy leaves of Acacia aroma plants were collected in the Chaco
Biogeographical Region in Northwest Argentina (Table 1 and Fig 1) and maintained in bags with
silica gel prior to DNA extraction. A total of six sites were sampled, each with different levels
of disturbance. No specific permissions were required in the field locations where samples
were taken, as they did not involve endangered or protected species and were not on private
lands. The number of sampled individuals ranged from 20 to 50 adults per site. Of these
individuals, 170 (95%) yielded clear amplifications allowing for population analysis (Table 1). The
sampling strategy used here covered between 70 and 100% of the adult individuals at each site,
due to the small size of the populations. The population of Lavalle (LA) is the most degraded
from anthropogenic activities, and currently, almost all of the area is deforested; QuimilÂõ (QU)
is located in a highly disturbed area, and its distribution is patchy; Tapia (TA) is a remnant of
only 20 individuals, surrounded by soy plantations; Robles (RO) extends on both sides of a
road impacted by a high level of traffic on this route; and San JoseÂ (SJ) and Mili (MI) are the
most pristine populations among the sampled sites.
For each sampled tree, the spatial coordinates (altitude, latitude, and longitude) were
recorded using a GPS device (Garmim1 eTrex).
The representative vouchers of each sampled tree were deposited at the SI Herbarium,
Instituto de BotaÂnica Darwinion, San Isidro, Buenos Aires, Argentina.
AFLP methods and data analysis
DNA extraction. The DNA of young leaves was extracted with the DNeasy Plant Kit
(QIAGEN Inc., Valencia, California, USA), following the manufacturer's instructions. The
DNA was stored at -20ÊC.
The AFLP assay was performed as described by Vos et al. [
], following the steps detailed
in Pometti et al. [
]. Specifically, for the present work on selective amplification, four primer
combinations were chosen: E + AAC/M + CTT (C2), E + AGG/M + CAG (C3), E + AAC/M +
CAA (C4) and E +AAG/M + CAA (C5). In all cases, primers M + 3 were labelled with the
fluorescent dye 6-FAM.
Data scoring and analysis. Each AFLP band was considered a single biallelic locus with
an amplifiable (dominant) allele, scored (1), and a null (recessive) allele, scored (0).
The Bayesian likelihood method implemented in the program BAYESCAN v2.1 [
utilized to identify the existence of variants with non-neutral divergence among the six
populations. The burn-in period was 50,000, the thinning interval was 10, the number of iterations
was 100,000, the number of pilot runs was 20 and the length of each pilot run was 5,000.
3 / 14
Fig 1. Map of Argentina showing the region and amplified zones where populations of A. aroma were sampled.
Population abbreviations: LA: Lavalle, MI: Mili, QU: QuimilÂõ, RO: Robles, SJ: San JoseÂ, TA: Tapia. Base maps were obtained
from http://viewer.nationalmap.gov/viewer/ and http://eros.usgs.gov/.
Therefore, all genetic diversity and population structure analyses were carried out on the set of
putatively neutral loci.
To estimate the allele frequencies, the software AFLP-SURV [
] was employed using the
Bayesian method with non-uniform prior distribution of allele frequencies, as described by
], following Lynch and Milligan's [
] approach. Nei's [
] genetic diversity
HE and pairwise Nei's [
] genetic distances between populations were also estimated using
the software AFLP-SURV [
4 / 14
To estimate the distribution of genetic diversity, the analysis of molecular variance
(AMOVA) was assessed by considering within-population and between-population
components. The decomposition of variance by AMOVA was conducted following Excoffier et al.
] while using the matrix approximations from Dyer et al. [
] with the software
]. Non-hierarchical Wright's  FST, was estimated with the package HierFstat
]. The significance of this estimate was obtained through a G test based on 5000
To assess isolation by distance, a Mantel test was performed using the function ªmantel.
randtestº from the ade4 package [
], by testing the relationship between pairwise NeiÂs
 and geographic distances.
To identify the population genetic structure in the populations of A. aroma, two Bayesian
model-based cluster analyses were contrasted: STRUCTURE [
] and Geneland [
rationale for comparing these methods is that they might yield different results because they
use different information (the former does not use geographical coordinates, whereas the latter
does) and produce different visualizations of the putative distribution of the clusters detected.
First, STRUCTURE version 2.3.4 [
] was used with a burn-in period and Markov chain
Monte Carlo (MCMC) repetitions set to 50,000 and 100,000, respectively. The admixture
model with correlated allele frequencies was selected; K was set at 1±8, and K values were
averaged across 10 iterations. Using the rate of change in the log likelihood, the ad hoc statistic ΔK
described by Evanno et al. [
] was estimated using STRUCTURE HARVESTER software
]. The results from STRUCTURE were edited with the software CLUMPP 1.1.2 [
Distruct 1.1 [
] to obtain the plot.
The second method relied on the spatial cluster model implemented in the Geneland
] of the program R [
]. Following the user's manual recommendations, the Markov
chain Monte Carlo (MCMC) repetitions were set to 100,000, the thinning was set to 100 and
the burn-in period was set to 100 (we eliminated the first 100 iterations when the curve was
not constant); the number of groups (K) to be tested was set from 1±7. Each individual was
assigned to one of K populations (1 K 7) based on its multilocus genotypes and spatial
coordinates. To confirm that the run was long enough, we obtained 10 different runs and
compared the parameter estimates (K, individual population membership, maps). The best result
was chosen based on the highest average posterior density.
To analyse the fine-scale spatial genetic structure (SGS), the approach described by Hardy
] was utilized, studying each population through kinship coefficients (Fij). The assumed
inbreeding coefficient was FIS = 0.19, based on the average of FIS estimates obtained for the
same species from codominant allozyme markers [
]. The number of distance classes
(distance intervals within which all pairs of sampling points are considered) was set to between 5
and 30 per population, in order to include at least 40 pairs of individuals in each distance class.
To establish the relationship between geographic distance classes and genetic similarity, the
regression slope of the kinship coefficients on log-transformed distances (bF) was estimated.
To determine the statistical significance of F1 (the mean kinship coefficient between
individuals belonging to the first distance class) and the bF, the upper and lower bounds of the
95% confidence interval of Fij were used, which were defined after 10,000 permutations of
individuals within locations. The Sp statistic [
] was computed for each population based on
the regression slope of kinship coefficients, as Sp = −bF/(1−F1). The Sp statistic was expected to
summarize the intensity of SGS, allowing for a quantitative comparison among species and/or
]. All estimations of SGS were performed using the software SPAGeDi v1.5 [
An indirect estimation of gene flow from the SGS estimates was performed assuming an
equilibrium of isolation by distance in the fine-scale genetic structure. In such cases, the extent
of gene flow can be expressed in terms of Wright's neighbourhood size as Nb 4πDEσg2, where
5 / 14
DE is the effective population density and σg is the mean-squared parent-offspring distance
and can be estimated as the inverse of Sp [
] yielding Nb = (F1−1)/bF. In this study, we
estimated Nb and σg using the census density of populations and three predicted effective values
(1/2, 1/4 and 1/10 of the census density).
In this study, a total of 852 AFLP bands in the interval of 50±400 bp were generated with the
four primer pair combinations used. From these bands, 401 with total reproducibility were
selected for all analyses to yield a 0% error rate.
The scan for FST-outliers, conducted with BAYESCAN within the 401 AFLP loci (with a
qvalue threshold of 10%), did not detect selection for any locus.
The measurements of genetic diversity are summarized in Table 1. HE varied from 0.18 in
TA to 0.24 in SJ (mean HE = 0.21), and PPL varied from 56.4% in TA to 67.1% in QU (mean
PPL = 62.1%).
When analysing the components of genetic diversity with AFLP-SURV, the highest
component was within populations (Hw = 0.21), while the between populations component
(Hb = 0.07) was lower. The analysis of population structure by means of Wright's  FST
statistic (0.42) was highly significant (P = 5x10-4). When comparing pairwise NeiÂs [
geographical distances matrices with the Mantel test, this result was not significant (P = 0.38).
The analysis of molecular variance indicated that the largest component of genetic diversity
(60.7%) was found within populations and that the remaining (39.3%) was found between
populations (Table 2).
Using the software STRUCTURE, a high peak of ΔK was found at K = 3, based on the AFLP
dataset analysis determining the presence of three clusters. In this analysis, individuals from
MI, RO, SJ and TA were grouped, as they were similar from a genetic perspective.
The remaining two sample sites of A. aroma, LA and QU, constituted two differentiated
clusters. Admixture was observed in all populations but was less prevalent in MI and SJ
Fig 2. Clustering of individuals produced by STRUCTURE for K = 3. Each individual is represented by a vertical coloured line. The same colour for
different individuals indicates that they belong to the same cluster or group. The population codes are the same as those shown in Fig 1.
6 / 14
Analysis using Geneland yielded a modal number of populations between 3 and 4, with a
higher proportion of K = 4 (Table 3). The run with the highest average posterior density was
selected. Clusters 1 and 2 included 33 and 10 individuals, respectively, from QU (Fig 3a and
3b). Cluster 3 was constituted by all individuals from LA and the remaining 7 individuals from
QU (Fig 3c); and cluster 4 was composed of individuals from TA, RO, SJ and MI (Fig 3d).
Fig 3. Spatial distribution of each group defined by Geneland at K = 4. (a) Cluster 1, (b) Cluster 2, (c) Cluster 3, and (d) Cluster 4. Clusters are
indicated by areas with different intensities of colour. Lighter-coloured areas indicate a higher probability that individuals belong to that cluster. The
population codes are the same as those shown in Fig 1.
7 / 14
Fig 4. Correlograms showing the fine-scale spatial genetic structure (SGS) of the studied populations. (a) Lavalle,
(b) Robles, (c) QuimilÂõ, (d) Mili, (e) San JoseÂ, (f) Tapia, (g) QuimilÂõ A, and (h) QuimilÂõ B. The filled circles indicate
significant Fij values. The dotted lines indicate the 95% confidence intervals (10,000 permutations).
Significant SGS was detected in short to medium distances in the LA, RO and QU
populations (up to 530 m) (Fig 4a, 4b and 4c, respectively). However, no significant spatial genetic
structure was detected in MI, SJ and TA (Fig 4d, 4e and 4f, respectively). In distance classes
shorter than 1000 m, a pattern of positive Fij was observed in almost all populations, and for
larger distance classes, negative Fij values were observed.
Populations QU and RO showed the strongest SGS and have a negative log slope (bF)
8 / 14
The neighbourhood size was calculated for populations where SGS was significant and
ranged from 15.2 in QU to 64.3 individuals in LA (Table 4).
The estimation of gene dispersal (σg) was conducted considering four different effective
densities in populations where SGS was significant. The extreme values of gene dispersal,
which corresponded respectively to the highest and lowest density estimates in each
population, were 273 and 864 m in LA, 45 and 141 m in RO, and 67 and 211 m in QU (Table 4).
Considering its patchy distribution, QuimilÂõ (QU) was split into two groups of individuals:
QUA (36 individuals) and QUB (14 individuals) (Fig 1). Significant SGS was detected in both
groups (Table 3, Fig 4g and 4h), but differences were observed between the groups in both
neighbourhood size and Sp estimates. The estimation of gene dispersal (σg) ranged from 86
to 273 m in QUA and 41 to 131 m in QUB, depending on the value used for the estimated
Due to the importance of the Acacia aroma species in South America, in this study, we
explored genetic diversity and population structure and characterized fine-scale spatial genetic
structure of this species in Argentina. In this work, we observed high levels of genetic diversity,
showing that populations of this species tend to maintain the majority of variability within
populations, as do other species of Acacia and perennial woody outcrossed species [
13, 37, 38,
]. Additionally, significant SGS was detected in 3 of 6 populations.
Habitat fragmentation is a significant threat to the maintenance of biodiversity in many
ecosystems. In general, the genetic consequences of habitat degradation focus on the reduced
size of populations and increased spatial isolation of remnant populations. However, in some
circumstances, fragmentation appears to increase gene flow among remnant populations,
breaking down local genetic structure [
]. Therefore, it is important to study the genetic
structure of populations when habitat degradation is occurring, in order to design management
plans to maintain the genetic diversity of a species. The genetic diversity observed in A. aroma
seemed to be relatively high when compared with other species sharing the same life history
traits (citar Nybom 2004) and was similar to the values observed in other Acacia species such
as A. curvifructa [
], A. visco [
] and A. senegal [
], which were also studied with dominant
markers. Additionally, in this study, the determination of population structure was assessed
with several approaches. WrightÂs FST statistics (0.42) showed a strong signal of genetic
structure among populations. However, the lack of significant correlation between geographic and
genetic distance suggested that the differentiation among populations did not fit the isolation
9 / 14
by distance model. This may be explained by assuming that gene flow is limited and that
populations have not reached the migration-drift equilibrium [
]. This approach was consistent
with the AMOVA results. Literature has established that long-lived and outcrossing species
tend to maintain most of their variability within populations [
]. This is the case for A.
aroma, that in this work showed that the majority of its molecular variance occurred within
populations. This result also concurs with previous records for other South American and
African species of Acacia [
40, 43, 44, 45,
The analysis with STRUCTURE showed that the optimal number of clusters (K) was 3.
Conversely, Geneland analysis showed that the optimal number of clusters was K = 4, which
shared a similar grouping to that of STRUCTURE but divided in two one of the clusters
identified by the aforementioned program. In the two cases, populations of SJ, RO, MI and TA
constituted a group. This finding could partially be due to the geographical proximity of SJ, RO
and MI. Although TA is geographically farther in distance, there is a route (used by vehicles
and pedestrians) that might promote gene flow (mediated by human activities) between TA
and the first three populations, leading to the increased genetic similarity. The population of
LA constituted a second cluster in STRUCTURE, with low levels of admixture. The individuals
from QU constituted a third cluster, with evidence of admixture with a sizeable contribution
from LA. This situation, with several differences, seems to be reflected in the plot from
Geneland (Fig 3). In this case, the admixed individuals in QU were clustered with those from LA.
The remaining individuals from QU were split into two clusters by Geneland.
Huang et al. [
] theorized that admixture, reflecting allele sharing, can result from
incomplete lineage sorting of historically contiguous populations, which might be the case for LA and
QU populations. The heterogeneity in the composition of QU, shown by Geneland, could be
because this population occurs in a highly disturbed area, and the distribution of its individuals
is patchy. Indeed, fragmentation and perturbation are expected to reduce the effective
population size within patches and increase the genetic differentiation between the populations [
Vekemans and Hardy [
] noted that individuals located near each other tend to be more
similar than those located farther apart. Additionally, isolation by distance modelling suggests
that SGS is expected at the equilibrium between drift and dispersal. Therefore, limited gene
dispersal may produce mating among related individuals and fine-scale spatial genetic
structure. In the present study, the individuals who were geographically farther apart were revealed
to have lower genetic similarity. Although the extent of gene dispersal (σg) seems to be larger
than the sampling area within sites, significant SGS was revealed in some populations,
suggesting an isolation by distance pattern.
Previous research has demonstrated that SGS is correlated with the mating system, life
history and population density [
]. Thus, the Sp statistic was proposed to synthesize SGS intensity
and remains useful when comparing the strength of SGS in different populations. A. aroma is
an outcrossing species, with its pollen and seed dispersed by animals. However, the mean Sp
statistic estimated here (Sp = 0.057) was greater than the mean values presented by Vekemans
and Hardy [
] for outcrossing species (Sp = 0.0126), species with animal-dispersed seeds
(Sp = 0.0088), and species with pollen dispersed through animals (Sp = 0.0171). Additionally,
this statistic was higher than the values observed for Schinus molle (Sp = 0.021) [
Prosopis alba (Sp = 0.003) [
] which are both species that share similar life history traits with A.
aroma, since they are outcrosser trees with animals as the vectors responsible for seed and
pollen dispersal. Spatial genetic structure can be affected by density, among other factors, because
it can influence the rate of genetic drift [
]. In support of this concept, the three populations
with the lowest densities, LA, RO and QU, are those which showed significant SGS. When
splitting QU into two groups, thereby noting its patchy distribution, level of disturbance,
geographical position of individuals and results obtained with Geneland, we observed a
10 / 14
remarkably higher Sp value in QUB (0.23) than in the QU sample as a whole (0.07) and in
QUA (0.03). The differences observed between the two groups could be partially attributed to
the low level of disturbance in QUB. This group of individuals, although small, is situated in a
more pristine and unmanaged area than those from QUA. Since the six populations studied
here presented different levels of disturbance and ecological characteristics, our work suggests
the need to evaluate the ecological aspects of the life history and landscape for each population.
When comparing the estimates of gene dispersal, we obtained similar values for the four
possible densities considered here as those estimated with AFLP markers in other outcrossers,
animal, pollen and seed dispersed trees, like Chrysophyllum sanguinolentum, Eperua
grandiflora and Virola michelii [
]. Our results on high gene dispersal are consistent with the basis
that A. aroma is a species pollinated by bees, with seeds dispersed by livestock and large
mammals. Additionally, the distance of gene dispersal is usually inversely related to population
], and this occurred in LA, RO and QU, where the density was lower than 9 individuals
per ha. Curtu et al. [
] proposed that low density can act as a barrier to pollen and seed
dispersal, thereby yielding SGS.
When aiming to design conservation, management, and sustainable use strategies for a species,
it is important to understand the patterns of genetic diversity [
42, 51, 52
]. In the Chaco Region
of Northwest Argentina, the contributing factors to environmental degradation have been
deforestation, uncontrolled firewood harvesting, livestock overstocking, and in some areas,
tillage of non-arable lands [
]. Analysis of population genetic structures shows that managers
should consider either three or four (depending on analysis) genetically distinct groups when
making management decisions. Our work describes, for the first time, the SGS and gene
dispersal parameters in A. aroma, which are valuable for the management and conservation of
this and other Acacia species. The current SGS analysis provides information that may be used
during sampling of individuals and seed collection for ex situ conservation and reforestation
programmes. Our findings suggest that for A. aroma reforestation and management
programmes, sampling should consider a minimal distance of 50 to 150 m to minimize genetic
relatedness among sampled seeds in each area. In cases of disturbed populations with low
density such as LA, the minimum and maximum distances should be much greater (270
and 870 m, respectively). In summary, the findings of this study are important for managing
and conserving the extant trees and populations of A. aroma in the fragmented landscapes,
such as Lavalle and QuimilÂõ, and provide baseline information on the spatial structuring and
dispersal of genes in this species. The combination of molecular markers and robust statistical
approaches constitutes an effective strategy for supporting programmes that mitigate the
deterioration and desertification of semi-arid lands in Argentina.
S1 Table. Allele frequencies for the 401 AFLP loci analysed in Acacia aroma populations.
Sample size (N); frequency of the AFLP fragment or marker (freq_frag); estimated frequency
of the null allele (freq_-all); estimated variance of the frequency of the null allele (var_-all).
The authors of this work want to thank the SERVICIO DE SECUENCIACIOÂ N Y
GENOTIPIFICADO from the Departamento de EcologÂõa GeneÂtica y EvolucioÂn, FCEyN, UBA. This work
11 / 14
was supported by the Universidad de Buenos Aires (UBA 20020130100043BA to B.O.S.) and
the Agencia Nacional de PromocioÂn CientÂõfica y TecnoloÂgica (PICTO 2011±0081 OTNA to
B.O.S., PICT-2013-0478 to J.C.V. and PICT-2013-1039 to C.L.P). We wish to thank the AJE
editors for the revisions to the language of our manuscript. Additionally, we acknowledge
the contributions of the three reviewers and the Associate Editor of PLOS ONE, who helped
improve the quality of our work.
Conceptualization: Carolina Pometti, Beatriz Saidman, Juan Vilardi.
Formal analysis: Carolina Pometti, Cecilia Bessega, Juan Vilardi.
Funding acquisition: Carolina Pometti, Beatriz Saidman, Juan Vilardi.
Investigation: Carolina Pometti.
Methodology: Carolina Pometti, Cecilia Bessega, Juan Vilardi.
Resources: Carolina Pometti, Cecilia Bessega, Ana Cialdella, Mauricio Ewens.
Writing ± original draft: Carolina Pometti, Cecilia Bessega, Ana Cialdella, Mauricio Ewens,
Beatriz Saidman, Juan Vilardi.
12 / 14
13 / 14
1. Peakall R , Ruibal M , Lindenmayer DB . Spatial autocorrelation analysis offers new insights into gene flow in the Australian bush rat, Rattus fuscipes . Evolution . 2003 ; 57 : 1182 ±1195 PMID: 12836834
2. Moran EV , Clark JS . Estimating seed and pollen movement in a monoecious plant: a hierarchical Bayesian approach integrating genetic and ecological data . Mol Ecol . 2011 ; 20 : 1248 ±1262 https://doi. org/10.1111/j. 1365 - 294X . 2011 . 05019 . x PMID : 21332584
3. Vekemans X , Hardy OJ . New insights from fine-scale spatial genetic structure analyses in plant populations . Mol Ecol . 2004 ; 13 : 921 ±935 PMID: 15012766
4. Loveless MD , Hamrick JL . Ecological determinants of genetic structure in plant populations . Annual Rev Ecol Syst . 1984 ; 15 : 65 ± 95
5. Loiselle BA , Sork VL , Nason J , Graham C . Spatial genetic structure of a tropical understory shrub, Psychotria officinalis (Rubiaceae) . Am J Bot . 1995 ; 82 : 1420 ± 1425
6. Young A , Boyle T , Brown T. The population genetic consequences of habitat fragmentation for plants . Tree 10 : 413 ± 418 .
7. Lowe AJ , Boshier D , Ward M , Bacles CFE , Navarro C . Genetic resource impacts of habitat loss and degradation; reconciling empirical evidence and predicted theory for neotropical trees . Heredity . 2005 ; 95 : 255 ±273 https://doi.org/10.1038/sj.hdy. 6800725 PMID: 16094300
8. Rico-Arce ML . American species of Acacia (Leguminosae: Mimosoideae) . ComisioÂn Nacional para el conocimiento y Uso de la Biodiversidad (CONABIO) . 2007 .
9. Arroyo MTK . Breeding systems and pollination biology in Leguminosae . 1981 . Pp. 723 ±769 in Polhill R. M. & Raven P. H . (editors), Advances in Legume Systematics , Volume 1 . Royal Botanic Gardens , Kew.
10. Gutierrez J & Armesto JJ . El rol del Ganado en la dispersion de las semillas de Acacia caven (Leguminosae) . Ci. & Invest. Agric. (Chile) . 1981 ; 10 : 3 ± 8 .
11. FernaÂndez OA , Busso CA. Arid and semi-arid rangelands: two thirds of Argentina . Rala report. 1997; N0 200 , pp. 20 .
12. Vos R , Hogers R , Bleeker M , Reijans M , Lee T , Hornes M , et al. AFLP: a new technique for DNA fingerprinting . Nucl Acids Res . 1995 ; 23 , 4407 ± 4414 . PMID: 7501463
13. Pometti CL , Bessega CF , Vilardi JC , Ewens M , Saidman BO . Genetic variation in natural populations of Acacia visco (Fabaceae) belonging to two sub-regions of Argentina using AFLP . Pl Syst Evol . 2016 ; 302 : 901 ± 910 . https://doi.org/10.1007/s00606-016-1306-6
14. Foll M , Gaggiotti O. A genome-scan method to identify selected loci appropriate for both dominant and codominant markers: a Bayesian perspective . Genetics . 2008 ; 180 : 977 ± 993 . https://doi.org/10.1534/ genetics.108.092221 PMID: 18780740
15. Vekemans X . AFLP-SURV version 1.0. Distributed by the author . Laboratoire de GeÂneÂtique et Ecologie VeÂgeÂtale, UniversiteÂ Libre de Bruxelles, Belgium. 2002 .
16. Zhivotovsky LA . Estimating population structure in diploids with multilocus dominant DNA markers . Mol Ecol . 1999 ; 8 : 907 ± 913 . PMID: 10434412
17. Lynch M , Milligan BG . Analysis of population structure with RAPD markers . Mol Ecol . 1994 ; 3 : 91 ± 99 .
18. Nei M. Analysis of gene diversity in subdivided populations . Proc Natl Acad Sci . 1973 ; U.S.A 70 : 3321 ± 3323 . PMID: 4519626
19. Nei M. Estimation of average heterozygosity and genetic distance from a small number of individuals . Genetics . 1978 ; 89 : 583 ± 90 . PMID: 17248844
20. Excoffier L , Smouse P , Quattro J . Analysis of molecular variance inferred from metric distances among DNA haplotypes: Application to human mitochondrial DNA restriction data . Genetics . 1992 ; 131 : 479 ± 491 . PMID: 1644282
21. Dyer RJ , Westfall RW , Sork VL , Smouse PE . Two-Generation Analysis of Pollen Flow Across a Landscape V: A Stepwise Approach For Extracting Factors Contributing to Pollen Structure . Heredity . 2004 ; 92 ( 3 ): 204 ± 211 . https://doi.org/10.1038/sj.hdy. 6800397 PMID: 14722580
22. Dyer RJ . 2008 . GeneticStudio: Population genetic software that doesn't suck . http://dyerlab.bio.vcu. edu.
Wright S. Evolution and the Genetics of Populations variability within and among natural populations , vol. 4 . University of Chicago Press, Chicago. 1978 .
24. Goudet J. hierfstat: Estimation and Tests of Hierarchical F-statistics . R Package Version 0 .04± 14 . 2014 .
25. R Development Core Team . 2016 . R: a language and environment for statistical computing . R Foundation for Statistical Computing: Vienna, Austria. ISBN 3-900051-07-0 . http://www.R-project.org
26. Dray S , Dufour AB . The ade4 package: implementing the duality diagram for ecologists . J Statistical Softw . 2007 ; 22 ( 4 ): 1± 20 .
27. Pritchard JK , Wen X , Falush D . STRUCTURE ver. 2.3 . 2009 . University of Chicago, Chicago, USA. http://pritch.bsd.uchicago.edu/
28. Guillot G , Mortier F , Estoup A. GENELAND: a computer package for landscape genetics . Mol Ecol Notes . 2005 ; 5 : 712 ± 715 .
29. Evanno G , Regnaut S , Goudet J . Detecting the number of clusters of individuals using the software STRUCTURE: a simulation study . Mol Ecol . 2005 ; 14 : 2611 ± 2620 . https://doi.org/10.1111/j. 1365 - 294X . 2005 . 02553 . x PMID : 15969739
30. Earl DA , vonHoldt BM . STRUCTURE HARVESTER: a website and program for visualizing STRUCTURE output and implementing the Evanno method . Conservation Genet Resour . 2012 ; 4 : 359 . https:// doi.org/10.1007/s12686-011-9548-7
31. Jakobsson M , Rosenberg NA . CLUMPP: a cluster matching and permutation program for dealing with label switching and multimodality in analysis of population structure . Bioinformatics . 2007 ; 23 : 1801 ± 1806 . https://doi.org/10.1093/bioinformatics/btm233 PMID: 17485429
32. Rosenberg NA . Distruct: a program for the graphical display of population structure . Mol Ecol Notes . 2004 ; 4 : 137 ± 138 .
33. Hardy OJ . Estimation of pairwise relatedness between individuals and characterization of isolation-bydistance processes using dominant genetic markers . Mol Ecol . 2003 ; 12 : 1577 ± 1588 . PMID: 12755885
34. Casiva PV , Vilardi JC , Cialdella AM , Saidman BO . Mating system and population structure of Acacia aroma and A. macracantha (Fabacecae) . Am J Bot . 2004 ; 91 ( 1 ): 58 ± 64 . https://doi.org/10.3732/ajb.91. 1.58 PMID: 21653363
35. Hardy OJ , Vekemans X . Spagedi: a versatile computer program to analyse spatial genetic structure at the individual or population levels . Mol Ecol Notes . 2002 ; 2 : 618 ± 620 .
36. Rousset F. Genetic differentiation between individuals . J Evol Biol . 2000 ; 13 : 58 ± 62 .
37. Pometti CL , Bessega CF , Vilardi JC , Cialdella AM , Saidman BO . Genetic diversity within and among two Argentinean and one Mexican species of Acacia (Fabaceae) . Bot J Linn Soc . 2015 ; 177 : 593 ± 606 . https://doi.org/10.1111/boj.12262
38. Nybom H . Comparison of different nuclear DNA markers for estimating intraspecific genetic diversity in plants . Mol Ecol . 2004 ; 13 : 1143 ± 1155 . https://doi.org/10.1111/j. 1365 - 294X . 2004 . 02141 . x PMID : 15078452
39. Pometti CL , Bessega CF , Vilardi JC , Saidman BO . Landscape genetic structure of natural populations of Acacia caven in Argentina . Tree Genet Genomes . 2012 ; 8 : 911 ± 924 . https://doi.org/10.1007/ s11295-012-0479-6
40. Chiveu CJ , Dangasuk OG , Omunyin ME , Wachira FN . Genetic diversity in Kenyan populations of Acacia senegal (L.) willd revealed by combined RAPD and ISSR markers . African J Biotech . 2008 ; 7 : 2333 ± 2340 .
41. Hedrick PW . Genetic of Populations. 3rd edition; Jones and Bartlett Publishers, Sudbury, Massachusetts. 2005 . Chapter 9, pp. 505 ± 507 .
42. Hamrick JL , Godt MJW . Allozyme diversity in plant species . In: Brown AHD , Clegg MT , Kahler AL , Weir BS , eds. Plant population genetics, breeding and genetic resourses. Sunderland , MA: Sinauer Associates, 43 ± 63 . 1989 .
43. Omondi SF , Kireger E , Dangasuk OG , Chikamai B , Odee DW , Cavers S , Khasa DP . Genetic Diversity and Population Structure of Acacia Senegal (L) Willd . in Kenya. Tropical Plant Biology . 2010 ; 3 : 59 ± 70 .
44. Pometti CL , Vilardi JC , Saidman BO . Mating system parameters and genetic structure in Argentinean populations of Acacia caven (Leguminosae, Mimosoideae) . Pl Syst Evol . 2011 ; 292 : 25 ± 32 . https://doi. org/10.1007/s00606-010-0389-8
45. Pometti CL , Bessega CF , Vilardi JC , Saidman BO . Comparison of mating system parameters and genetic structure in three natural scenarios of Acacia visco (Leguminosae, Mimosoideae) . Pl Syst Evol . 2013 ; 299 : 761 ± 771 . https://doi.org/10.1007/s00606-013-0759-0
46. Huang CL , Chen JH , Tsang MH , Chung JD , Chang CT , Hwang SY ( 2015 ) Influences of environmental and spatial factors on genetic and epigenetic variations in Rhododendron oldhamii (Ericaceae) . Tree Genet Genomes 11 : 823 . https://doi.org/10.1007/s11295-014-0823-0
47. Lemos RPM , D'Oliveira CB , Stefenon VM ( 2015 ) Genetic structure and internal gene flow in populations of Schinus molle (Anacardiaceae) in the Brazilian Pampa . Tree Genet Genomes 11 : 75 . https://doi.org/ 10.1007/s11295-015-0885-7
48. Bessega C , Pometti CL , Ewens M , Saidman BO , Vilardi JC ( 2016 ) Fine-scale spatial genetic structure analysis in two Argentine populations of Prosopis alba (Mimosoideae) with different levels of ecological disturbance . Eur J Forest Res 135 : 495 ±505 https://doi.org/10.1007/s10342-016-0948-9
49. Hardy OJ , Maggia L , Bandou E , Breyne P , Caron H , Chevallier ME , Doligez A , Dutech C , Kremer A , Latouche-Halle C , Troispoux V , Veron V , Degen B . Fine-scale genetic structure and gene dispersal influences in 10 neotropical tree species . Mol Ecol . 2006 ; 15 : 559 ± 571 . https://doi.org/10.1111/j. 1365 - 294X . 2005 . 02785 . x PMID : 16448421
50. Curtu AL , Craciunesc I , Enescu CM , Vidalis A , Sofletea N. Finescale spatial genetic structure in a multioak-species (Quercus spp.) forest . iForest. 2015 ; 8 : 324 ± 332 . https://doi.org/10.3832/ifor1150- 007
51. Hamrick JL . The distribution of genetic variation within and among natural plant populations . In: Schonewald-Cox C.M. , Chambers S.M. , McBryde B. , Thomas W.L. (Eds.), Genetics and Conservation . Benjamin/Cummings, Menlo Park, California, USA, pp. 335 ± 348 . 1983 .
52. Francisco-Ortega J , Santos-Guerra A , Kim SC , Crawford DJ . Plant genetic diversity in the Canary Islands: a conservation perpective . Am J Bot . 2000 ; 87 : 909 ± 919 . PMID: 10898768