Content uploaded by Stefanie Reim
Author content
All content in this area was uploaded by Stefanie Reim on Jan 31, 2014
Content may be subject to copyright.
RESEARCH ARTICLE
Diversity of the European indigenous wild apple (Malus
sylvestris (L.) Mill.) in the East Ore Mountains
(Osterzgebirge), Germany: II. Genetic characterization
Stefanie Reim •Aki Ho
¨ltken •Monika Ho
¨fer
Received: 12 October 2011 / Accepted: 9 July 2012
ÓSpringer Science+Business Media Dordrecht 2012
Abstract In the present study genetic diversity and
hybridization with cultivars were investigated in a
population of the endangered European wild apple
species Malus sylvestris (L.) Mill. with the aim to
establish a basis for the implementation of conserva-
tion activities and to ensure its long-term preservation.
A total of 284 putative M. sylvestris trees located in the
East Ore Mountains were investigated along with a
standard set of reference apple genotypes proposed by
the European Cooperative Program for Plant Genetic
Resources (ECPGR) and 13 old apple cultivars often
cultivated in Saxony. The genetic analysis was
performed using 12 microsatellite markers also rec-
ommend by the ECPGR. To differentiate ‘true type’
M. sylvestris individuals, hybrids and apple cultivars
(Malus 9domestica Borkh.) a model-based cluster
analysis was performed using STRUCTURE. Two
clusters were identified consisting of M. sylvestris and
M. 9domestica genotypes. About 40 % of the puta-
tive M. sylvestris showed an admixture of the species-
specific allele frequencies and were defined as hybrids.
The genetic diversity of the ‘true type’ M. sylvestris
population was still high but slightly lower than in the
apple cultivars especially since some SSR loci were
fixed on one or few alleles in the M. sylvestris
population. The differentiation parameters between
‘true type’ wild apple and cultivars indicated a clear
discrimination between the wild and cultivated apple
individuals. This fact confirms our expectation of the
existence of ‘true type’ M. sylvestris individuals in the
East Ore Mountains and argues for the realization of
preservation measures in this area.
Keywords Conservation Hybridization Malus
sylvestris Microsatellites
Abbreviations
ECPGR European Cooperative Program for Plant
Genetic Resources
HWE Hardy–Weinberg equilibrium
Introduction
Malus sylvestris (L.) Mill. is the only indigenous wild
apple species in Central Europe and distributed from
South-Scandinavia to the Iberian Peninsula and from
the Volga to the British Isles (Robinson et al. 2001).
S. Reim M. Ho
¨fer (&)
Institute for Breeding Research on Horticultural and Fruit
Crops, Julius Ku
¨hn-Institute, Pillnitzer Platz 3a, 01326
Dresden, Germany
e-mail: monika.hoefer@jki.bund.de
A. Ho
¨ltken
Department of Wood Technology and Wood Biology,
University of Hamburg, Leuschner Straße 91, 21031
Hamburg, Germany
A. Ho
¨ltken
Plant Genetic Diagnostics GmbH, Alte Landstraße 26,
22927 Großhansdorf, Germany
123
Genet Resour Crop Evol
DOI 10.1007/s10722-012-9885-8
Despite this large range, the European wild apple is a
rare and endangered species occurring as single
individuals or small and scattered groups of trees
(Kleinschmit and Stephan 1997). Its economical
insignificance and inconspicuous appearance in com-
bination with present-day intensive forest and agri-
cultural management systems induced a further
decrease of this species. As a result, fragmentation
and spatial isolation of suitable habitats may have led
to a reproductive separation of long-established
metapopulation networks, even in species adapted to
naturally scattered distributions and small population
sizes (Wagner 2006).
Due to its gametophytic self-incompatibility sys-
tem Malus species are dependent on compatible
pollination partners to maintain natural regeneration
(Broothaerts 2003). Apple pollen is mainly distributed
by vectors and no reproductive barriers are existing
and inhibiting hybridization of different species within
the genus Malus (Larsen and Kjaer 2009; Reim et al.
2006). Thus, due to the low availability of potential
wild apple pollination partners we can expect a high
probability of admixture with domesticated apple
(M. 9domestica Borkh.) (Rosenthal 2003; Larsen
et al. 2008). As a result, loss of genetic identity of M.
sylvestris and replacement by hybrids reducing the
viability of wild apple populations may have increased
the risk of species extinction (Allendorf et al. 2001;
Rhymer and Simberloff 1996; Rieseberg and Ellstrand
1993).
An essential prerequisite for the development of
sustainable conservation strategies of M. sylvestris is
knowledge about population genetic structures and the
ability to discriminate between ‘true type’ M. sylvestris
genotypes and existing hybrids. This is of particular
relevance in order to identify important genetic
resources, to remove potential hybrids and, as primary
objective, to ensure the genetic integrity of M. sylves-
tris populations (Castiglione et al. 2010; Dreesen et al.
2010; Paul et al. 2000). For the classification of M.
sylvestris a number of morphological characteristics
are available from which hairiness of leaves and
flowers, fruit size and cover color are assumed to be the
most suitable (Fellenberg 2001; Remmy and Gruber
1993; Wagner 1996). The disadvantage of morpho-
logical traits is their variation depending on various
environmental influences. Particularly, leaf and flower
hairiness serve as a protection against evaporation
and vary during the growing season. Further, the
classification of the crab apple on basis of fruit size is
also not ensured since trees growing in forests in many
cases never bloom and bear fruits. Therefore, genetic
markers have increasingly been applied in recent years
for taxonomical studies of species. In apple species
specific microsatellite markers (Simple Sequence
Repeats; SSRs) have become the markers of choice
for the study of genetic diversity and provide an
accurate evaluation of relationships within the genus
Malus (Cavanna et al. 2008; Coart et al. 2003;
Gharghani et al. 2009; Guarino et al. 2006). SSR
markers are also used as a tool to indentify ‘hybrids’
within potential wild apple populations (Coart et al.
2003; Koopman et al. 2007; Larsen et al. 2006).
The aim of our study was to identify ‘true type’
individuals as well as hybrids within a M. sylvestris
population in the East Ore Mountains in Germany and
to investigate their genetic structures. This region
represents a well suited study area due to good
growing conditions for European wild apple and a
fairly large existing stock (Kleinschmit 1998; Natzke
and Pech 1998; Patzak 2003). Furthermore, a low
hybridization in the past between M. sylvestris and
their domesticated relatives can be expected due to the
late settlement and, hence, a comparatively late
cultivation of apple cultivars in the East Ore Moun-
tains. These facts indicate for the existence of
numerous ‘true type’ M. sylvestris individuals in the
East Ore Mountains and argue for the implementation
of conservation measures in this area. On the basis of
these prerequisites the main objectives of this work are
(1) to estimate hybridization with apple cultivars on
basis of genetic data and (2) to examine the genetic
diversity within a ‘true type’ M. sylvestris population.
Materials and methods
284 M. sylvestris individuals of the East Ore Moun-
tains were selected based on numerous morphological
traits. From these, 154 individuals were classified as
‘true type’ M. sylvestris, and 130 individuals as
putative hybrids. But, for reasons of simplicity, all
samples are designated as putative M. sylvestris in the
following. Detailed information about the morpho-
logical evaluation is given in Reim et al. (2011).
Additionally, seven genotypes (Table 1) recom-
mended by the European Cooperative Program for
Plant Genetic Resources (ECPGR) as standard apple
Genet Resour Crop Evol
123
genotypes (Evans et al. 2007) and further 13 old apple
cultivars often cultivated in Saxony were analyzed as
out-group genotypes.
The leaf material was collected from the trees in the
East Ore Mountains in 2 ml reaction tubes and dried
using silica gel (according to a modified protocol by
Slotta et al. 2008). The leaf material was stored at
room temperature until DNA isolation. The dried leaf
material was ground in a Mixer mill apparatus
(Retsch, Germany) and DNA isolation followed
immediately using the DNeasy Plant Mini Kit
(Qiagen, Germany) according to the manufacturer’s
instructions. The DNA quantity was estimated com-
paring a dilution series of 10, 20, 30, 40 and 50 ng of
k—DNA using the Quantity OneÒsoftware of the gel
documentation system (Biorad, Germany). All sam-
ples were diluted to 10 ng/ll.
Microsatellite analysis
From a set of standard microsatellite primers, defined
by ECPGR, 12 primer pairs were selected for the
present study (Table 1) (Evans et al. 2007). All these
selected SSR markers have been developed in apple
cultivars (Hokanson et al. 1998; Liebhard et al. 2002;
Silfverberg-Dilworth et al. 2006).
The forward primers in one multiplex reaction were
labeled with three different dyes (D2: Dye 751,
absorption max. 751 nm; D3: BMN-6, absorption
max 681 nm: D4: BMN-5, absorption max 645 nm;
Biomers, Germany). The PCRs were carried out in four
multiplex reactions each with three primer pairs
following the manufactures guide of the ‘type-it
microsatellite kit’
Ò
(Qiagen, Germany). The multiplex
electrophoresis was performed on a CEQ 2000 Genetic
Analysis System and the data analyzed using CEQ
2000 software (both Beckman Coulter, Germany).
Genetic clustering and grouping of individuals
A model-based clustering method was applied for all
304 individuals to identify distinct (random mating)
populations, assigning the individuals to groups and
detecting potential gene flow between apple cultivars
and indigenous wild apple stocks by using the
software STRUCTURE ver. 2.3.3. (Pritchard et al.
2000). To estimate the optimal number of populations
(K) we run the program with Kvarying from 2 to 3
with 5 runs for each Kvalue. Our parameters were
40,000 burn-in periods and 40,000 Markov chain
Monte Carlo repetitions using the admixture model
with correlated allele models.
Table 1 Allele sizes (in basepairs) of the recommended SSR fingerprinting set for the 7 apple control genotypes (ECPGR)
(Evans et al. 2007)
SSR primer MP
a
lg
b
Delicious Fiesta Prima Worcester
Pearmain
Malus
floribunda ‘821’
Malus
robusta ‘5’
Rootstock
‘Malling 9’
CH01h10* 1 8 88:96 101:101 94:101 96:101 101:109 86:109 96:113
CH04c07* 1 14 118:133 108:113 106:108 108:110 108 106:109 106:114:129
CH01h01* 1 17 115 122:134 118:122 111:129 103:137 86:97 113:119
Hi02c07** 2 1 114:116 116:151 110:118 114:151 114:136 116:118 116
CH01f03b* 2 9 136:178 158:170 136:158 136:170 148 170 158:170
GD147*** 2 13 137:152 145:150 131:150 137:150 123 145:150 139:152
CH02d08* 3 11 210:216 224:254 254 210:250 214:218 210:212 212:254
CH04e05* 3 7 173:202 199:226 173:208 173:200 187:197 181 197:220
CH02c11* 3 10 205:231 215:227 227:231 221:225 221:225 203:217 213:233
CH01f02* 4 12 178:182 180:203 178:205 186:205 174:178 174:178 168:170
CH02c09* 4 1 244:254 232:248 232:242 232:244 230:250 247 244
GD12*** 4 3 147:153 148 182:190 148 148:172 150:151 148:160
* Liebhard et al. (2002); ** Silfverberg-Dilworth et al. 2006; *** Hokanson et al. 1998
a
Adapted multiplex combinations for the usage of CEQ 8000 (Beckman Coulter Inc., Fullerton, CA)
b
Linkage group
Genet Resour Crop Evol
123
Comparing morphological and genetic data
To investigate correlations between genetic and mor-
phological distances a Mantel test with statistical
testing by 9,999 permutations were conducted by using
GENALEX ver. 6.3. (Peakall and Smouse 2006).
Therefore to perform the Mantel test, firstly two
distance matrices based on the genetic and morpho-
logical data were calculated also using GENALEX ver.
6.3. For the calculation of the morphological distance
matrix, we considered each morphological trait score
to be analogous to an allele at a haploid locus. The
Mantel correlation coefficient (rxy) ranging from -1to
1, an rxy of 0 indicates no relationship and an rxy close
to 1 indicates a strong positive relationship between
both distance matrices.
Additionally, the relationship between the admix-
ture coefficient (Q) calculated on the basis of the
genetic data using STRUCTURE and each single
morphological descriptor were compared by calculat-
ing Spearman’s rank correlation using the statistic
software SAS ver. 9.2. (Spearman 1904).
Genetic diversity parameters
The following genetic parameters within the complete
M. sylvestris collection, the estimated hybrids, the
estimated ‘true type’ M. sylvestris population and the
apple cultivars were calculated using the software
GENALEX ver. 6.3: mean number of alleles by locus
(N
a
), effective number of alleles (N
e
), number of
private alleles with frequencies over 5 % (pa),
observed heterozygosity (H
o
) and the expected heter-
ozygosity (H
e
) according to Nei (1972).
Test of deviations from Hardy–Weinberg equilib-
rium (HWE) within the complete M. sylvestris
collection, the hybrids and the ‘true type’ M. sylvestris
population was performed using Chi-squared statistics
v
2
(Hedrick 2005) calculated by the software program
GENALEX ver. 6.3. In this context we also deter-
mined the inbreeding coefficient F
is
to detect homo-
zygote or heterozygote excess, respectively, according
to Wright (1951):
Fis ¼HeH0
He
The F
is
value defines the degree of genetic
inbreeding within the populations due to non random
mating and can range from -1 (all individuals
heterozygous) to ?1 (no observed heterozygous).
Genetic fixation and differentiation
To measure divergence in population genetic struc-
tures between apple cultivars and their European wild
relatives, we calculated the genetic fixation index F
st
(Wright 1965) using the software program GENALEX
ver. 6.3 as well as genetic distance according to
Gregorius (1974) and Nei (1972). F
st
measures
relative genetic differentiation in the sense of the
extent to which the process of fixation has gone toward
completion but not in the sense implied in the extreme
case by absence of any common allele. Therefore, F
st
reaches its maximum value (F
st
=1) only if all
populations are monomorphic but not fixed for the
same allele (Wright 1978;Ho
¨ltken et al. 2003).
The significance of each variance component was
assessed using a permutation test (999 permutations).
The potential effect of null alleles on genetic differ-
entiation was calculated using the excluding null allele
method by the software program FREENA (Chapuis
and Estoup 2007). The estimation of null allele
frequencies on each locus should clarify a potential
deviation from Hardy–Weinberg equilibrium.
The genetic distance (D) (Gregorius 1974) between
the complete M. sylvestris collection and M. 9domes-
tica individuals as well between the ‘true type’ M.
sylvestris population and M. 9domestica individuals
was also calculated using the software GDA_NT
(Degen 2008). This parameter is an absolute measure
of differentiation, measuring the proportion of genetic
types not shared by both of the populations. It reaches
its maximum value 1, if the two populations have no
genetic types in common and its minimum value 0 if
the two populations have identical genetic structures.
For comparison with other studies Nei’s (1972)
genetic distance (D
s
) was also calculated using the
software POPGENE ver. 1.32 (Yeh et al. 2000).
Results
Range of allele sizes and allelic profile
In order to standardize international protocols, a
primer set of 12 SSRs based on the recommendations
of the ECPGR was chosen for this study. The selection
Genet Resour Crop Evol
123
of this primer set is based on high polymorphism,
location of single loci on different linkage groups and
multiplexing ability, tested on apple cultivars
(Table 1). All 12 genomic SSR markers showed
reproducible results with one to two amplified frag-
ments in all 284 M. sylvestris genotypes.
The estimation of the range of allele sizes and the
allelic profile includes the ‘true type’ M. sylvestris
individuals only, since the consideration of the hybrids
may distort the results because of their shared cultivar
specific alleles. The allele size range was in eight cases
(CH01H10, CH04C07, CH01H01, Hi02C07, GD147,
CH04E05, CH02C11, CH02C09) considerably higher
in M. sylvestris than specified in the ECPGR geno-
types of apple cultivars (Fig. 1). At locus CH01F02 we
found the exact opposite, a smaller range of allele sizes
in M. sylvestris compared to the apple cultivars. The
remaining three SSR loci showed approximately equal
allele size ranges (Table 1).
Malus sylvestris private alleles were detected in
eight SSR loci (CH01H10, CH01H01, Hi02C07,
CH01F03b, GD147, CH02D08, CH01F02 and
GD12) with a total of 16 private alleles with frequen-
cies exceeding 5 % (Fig. 1).
Particularly in the M. sylvestris population, the
variability of some SSR loci was restricted to only a
few alleles. At SSR locus CH04E05, most of the
M. sylvestris individuals were fixed on the 176 bp
allele with a frequency of 96 %. Similar results were
obtained at the SSR locus CH04C07 with only two
alleles (108 and 96 bp) with frequencies of 62 and
24 % and at locus CH02C09 showing the 234 bp allele
with a frequency of 61 %.
Genetic admixture of the species and identification
of hybrids
The Bayesian cluster analysis was performed for 304
genotypes using the STRUCTURE software (Fig. 2).
The optimal number of populations (K=2) was
chosen on basis of the estimated likelihood of data as
recommended in the manual. As a result the model
assumed two clusters which are characterized by a set
of allele frequencies at each of the 12 SSR loci. On this
basis the individuals were assigned probabilistically
into M. sylvestris and cultivar clusters.
All 18 M. 9domestica cultivars and the two wild
genotypes M. floribunda Siebold clone ‘821’ and
M. 9robusta (Carr.) Rehd. clone ‘5’ were assigned in
the cultivar cluster with an admixture coefficient
Q[0.20. An admixture coefficient of Q[0.80 was
chosen to separate ‘true type’ individuals from hybrids
based on comparing the morphological dataset scored
within the complete M. sylvestris collection (data not
shown). The correlation between the classification
‘true type’ M. sylvestris and hybrids based on the
morphological and genetically data was the highest by
using a threshold Q[0.80.
In total 167 individuals were classified with
Q[0.80 as true type M. sylvestris and 99 individuals
with an admixture coefficient Q=0.20–0.79 as
hybrids (Fig. 2). For the remaining 18 individuals
sampled in nature as M. sylvestris a percental mem-
bership of Q\0.20 to the M. sylvestris cluster was
calculated assuming that the hybridization input of
these individuals are very high or these individuals are
even feral growing apple cultivars. Despite of the last
assumption, these individuals are assigned as ‘hybrid’
in the further analysis.
The Mantel test including the phenotypic
and genetic distance matrix yielded a significant
(P
(rxy-rand[rxy-data)
=0.0001) correlation coefficient
of rxy =0.39, indicating a moderate correlation of
these two distance matrices.
Regarding the relationship of the single morpho-
logical descriptors’ scores and the population assign-
ment based on the genetically data, 10 of 17
morphological traits were significantly correlated with
the admixture coefficient Q, whereby the single
features ‘flower hairiness’ showed the highest corre-
lation (Table 3).
Genetic variation
The genetic diversity was determined based on the
calculation of allele frequencies of the complete
M. sylvestris collection (284 individuals) and sepa-
rately for the estimated hybrids (117 individuals) and
the declared ‘true type’ M. sylvestris population (167
individuals) as well as for the 18 M. 9domestica
individuals. The parameters of the genetic variability
estimated for each SSR marker and all investigated
groups are shown in Table 2. The wild apple geno-
types ‘M. floribunda 821’ and ‘M. robusta 5’ were not
included in the calculation because they were repre-
sented by only one individual and therefore could not
form own populations.
Genet Resour Crop Evol
123
Genet Resour Crop Evol
123
The average number of alleles (N
a
) within the
complete putative M. sylvestris collection was
N
a
=19, approximately equal for the hybrids
N
a
=18 and much lower for the ‘true type’
M. sylvestris population with N
a
=13, with a range
between 7 and 19 of alleles per locus. The average
effective number of alleles (N
e
) was nearly similar for
all analyzed groups with the highest N
e
=7 within the
hybrids and the lowest N
e
=5, varying between one
and eight per locus within the ‘true type’ M. sylvestris
population.
Description of individuals: bar 1 – 167: putative ‚true type‘ M. sylvestris individuals
bar 168 -284: putative hybrids
bar 285-302: apple cultivars
bar 303: M. floribunda ‚821‘
bar 304: M. robusta ‚5‘
Fig. 2 Genetic admixture of the 304 sampled genotypes of wild apple (M. sylvestris) and the apple cultivars calculated for two clusters
(K=2) using the program STRUCTURE 2.3.2 based on 12 SSR markers. Y-axis: admixture coefficient Q
Fig. 1 Allele frequencies of 12 SSR loci recommended by the
ECPGR in ‘true type’ M. sylvestris individuals found in the East
Ore Mountains and the investigated apple cultivars
b
Genet Resour Crop Evol
123
Regarding single loci, the expected heterozygosity
(H
e
) was high for most of the SSR loci in the complete
M. sylvestris group as well as for the separated hybrid
and ‘true type’ group with exception of the locus
CH04E05 (complete M. sylvestris group H
e
=0.16;
hybrid group H
e
=0.26 and ‘true type’ M. sylvestris
group H
e
=0.08).
The average expected heterozygosity varied
between H
e
=0.71 and H
e
=0.79. As might be
expected, the hybrids showed with H
e
=0.79 the
highest value followed by the complete M. sylvestris
collection with H
e
=0.75. Within the ‘true type’
M. sylvestris population the average expected heter-
ozygosity was with H
e
=0.71, a little bit lower as the
Table 2 Genetic variability parameters estimated for each SSR marker both for wild apple (M. sylvestris) found in the East Ore
Mountains and investigated apple cultivars
Locus Complete M.sylvestris collection
a
‘True type’ M. sylvestris
b
Size range N
a
N
e
H
o
H
e
Size range N
a
N
e
H
o
H
e
CH01H10 82–148 25 7 0.84 0.86 86–148 19 6 0.84 0.82
CH04C07 88–142 23 3 0.69 0.66 96–136 12 2 0.59 0.56
CH01H01 95–145 24 10 0.82 0.90 107–145 17 8 0.82 0.88
Hi02C07 94–150 24 6 0.83 0.83 100–148 17 5 0.80 0.78
CH01F03b 134–180 16 5 0.72 0.78 140–180 10 4 0.72 0.73
GD147 119–157 18 7 0.79 0.85 133–157 11 5 0.74 0.81
CH02D08 209–257 18 4 0.60 0.77 209–257 8 4 0.57 0.73
CH04E05 166–220 12 1 0.16 0.16 176–220 7 1 0.08 0.08
CH02C11 208–242 18 9 0.77 0.89 210–242 16 7 0.73 0.87
CH01F02 160–224 23 8 0.86 0.88 160–224 19 7 0.90 0.86
CH02C09 206–258 15 3 0.60 0.63 206–258 12 2 0.59 0.60
GD12 141–185 16 4 0.72 0.76 141–185 9 4 0.75 0.75
Mean 19 6 0.70 0.75 13 5 0.68 0.71
Locus Hybrids
c
Cultivars
d
Size range N
a
N
e
H
o
H
e
Size range
e
N
a
N
e
H
o
H
e
CH01H10 82–148 24 9 0.84 0.89 94–114 9 3 0.83 0.71
CH04C07 88–142 23 4 0.82 0.77 98–135 10 6 0.94 0.83
CH01H01 95–145 24 11 0.82 0.91 114–134 8 6 0.89 0.83
Hi02C07 94–150 23 8 0.88 0.88 108–149 8 5 1.00 0.80
CH01F03b 134–180 15 6 0.73 0.83 139–183 8 5 0.83 0.81
GD147 119–157 18 9 0.85 0.89 135–155 9 5 0.89 0.81
CH02D08 209–257 18 6 0.66 0.82 210–254 10 7 0.83 0.85
CH04E05 166–220 10 1 0.25 0.26 174–227 11 5 0.94 0.81
CH02C11 208–240 17 10 0.84 0.90 219–239 13 10 0.83 0.90
CH01F02 160–224 20 10 0.82 0.90 153–227 12 9 0.78 0.89
CH02C09 233–258 13 3 0.64 0.68 233–257 8 6 1.00 0.84
GD12 141–185 16 4 0.69 0.76 150–200 8 3 0.56 0.62
Mean 18 7 0.74 0.79 10 6 0.86 0.81
N
a
number of different alleles, N
e
number of effective alleles (=1/(Pp
i
2
)), p
i
relative frequency of the ith allele, pa number of private
alleles with frequencies over 5 %, H
o
observed heterozygosity (=number of heterozygotes/N), H
e
expected heterozygosity (=1 -P
p
i
2
)
a
284 potential M. sylvestris Individuals;
b
167 ‘true type’ M. sylvestris Individuals;
c
117 Hybrids;
d
18 Cultivars (the wild apple
genotypes ‘M. floribunda 821’ and ‘M. robusta 5’ were not included in the calculation);
e
allele size range specifications refer to the
publication of Evans et al. (2007)
Genet Resour Crop Evol
123
two other groups but also indicating a high genetic
diversity of the population.
Compared to the 18 apple cultivars the average number
of detected alleles (N
a
) was higher in each group of the
M. sylvestris collection. That was to be expected because
of the larger sample size of each M. sylvestris group (284,
167 or 117 individuals versus 18 apple cultivars).
In contrast, the average number of effective alleles
(N
e
) was higher for the apple cultivars than in the ‘true
type’ M. sylvestris population indicating that some
alleles were present only in one or few ‘true type’
M. sylvestris individuals. The expected heterozygosity
H
e
in the cultivars was also higher whereas single loci
presented remarkable deviations in the value of H
e
compared to the three groups of the wild apple
collection (CH04E05 and CH02C09).
Within the total M. sylvestris collection at all SSR-
loci, except one (CH01H10), the probability of the
observed v
2-
values were statistically significant indicat-
ing also deviations from the Hardy–Weinberg equilib-
rium (HWE). Regarding the separated group of hybrids
the HWE was similar to the complete M. sylvestris
collection with significant deviation at ten SSR loci. In
contrast, within the ‘true type’ M. sylvestris population
at six SSR-loci (Hi02C07, CH01F03b, Ch02D08,
Ch04E05, CH01F02, CH02C09) no deviation from
HWE could be observed (Table 3). For the other six loci
(CH01H10, CH04C07, CH01H01, GD 147, CH02C11
and GD12) a significant deviation (P\0.001) from
HWE could also be observed.
Null alleles may cause a locus to deviate from HWE
wherefore potential null allele frequencies for each
SSR marker were estimated by using FREENA. As
result in both, the complete M. sylvestris collection
and in the hybrid group, for each locus the presence of
null alleles can be assumed. Within the ‘true type’ M.
sylvestris only in case of six out of the 12 SSR markers
tested the presence of null alleles cannot be excluded
(CH01H10, CH04C07, CH01H01, GD 147, CH02C11
and GD12) (Table 4).
As would be anticipated from the presence of null
alleles within the complete M. sylvestris collection and
the hybrids, the average F
is
values with 0.059 and
0.056 calculated by using GenAlex were relatively
high. In the present study an average F
is
value of 0.036
was calculated indicating a slight excess of homozy-
gous individuals within the populations.
Genetic differentiation between the species
The partition of diversity between each wild apple
group and the apple cultivars was performed
using Wright’s F-statistics (Wright 1965) (Table 4).
Between the complete M. sylvestris collection and the
apple cultivars, the mean F
st
value was 0.09 indicating
a remarkable genetic differentiation. Separating the M.
sylvestris collection in hybrids and ‘true type’ groups,
the mean F
st
value decreases to 0.063 for the hybrids.
In contrast, reasonably to be expected the mean F
st
value within the ‘true type’ samples increase to 0.117
indicating a distinct genetic differentiation between
the ‘true type’ population and the apple cultivars if the
degree of fixation on different alleles is concerned
(Wright 1978). This becomes particularly apparent at
the SSR locus CH04E05 (F
st
=0.564) with the ‘true
type’ M. sylvestris population nearly fixed on allele
176 and the apple cultivars with a high allelic variation
range from 176 to 230 bp. This tendency can also be
observed for the loci CH01H10, CH02C09 and
CH02D08.
After adjustment of the F
st
values using FREENA
by accounting for potential null alleles, the F
st
value of
the complete M. sylvestris collection and the ‘true
type’ wild apples (F
st (FREENA)
=0.088 or F
st (FREE-
NA)
=0.116) was only slightly lower indicating that
null alleles do not have substantial impact on this
fixation index. In contrast to this, the F
st
value of the
hybrids was more influenced due to potential null
alleles and showed after the adjustment by FREENA a
remarkable lower value (F
st (FREENA)
=0.054 versus
F
st
=0.063).
The quantification of genetic differentiation
(D) according to Gregorius (1974) between the
complete M. sylvestris collection and the apple
cultivars resulted in a mean value of D=0.481. After
the exclusion all individuals estimated as hybrids the
amount of genetic differentiation was determined to a
mean value of D=0.583 indicating that almost 60 %
of allelic types have to be exchanged between ‘true
type’ M. sylvestris individuals and the apple cultivars
in order to obtain identical genetic structures. In other
words, wild and cultivated apple only share about
40 % of the allelic variants. This high value of genetic
differentiation was confirmed by Nei’s genetic dis-
tance parameter (D
s
=0.572).
Genet Resour Crop Evol
123
Table 3 Deviations from Hardy-Weinberg-Equilibrium (HWE), estimated null allele frequencies and inbreeding coefficients F
is
calculated for the wild apple collection from the
East Ore Mountains in Germany; genetic differentiation between the M. sylvestris population and the apple cultivars calculated by F-statistics (Wright 1951), the modified
F
ST(FREENA)
excluding potential null alleles (Chapuis and Estoup 2007)
Locus Total M. sylvestris collection
a
‘True type’ M. sylvestris
b
Hybrids
c
HWE Null
alleles
frequency
F
is
F
st
F
st
(FREENA)
HWE Null
alleles
frequency
F
is
F
st
F
st
(FREENA)
HWE Null
alleles
frequency
F
is
F
st
F
st
(FREENA)
CH01H10 ns 0.024 0.020 0.131 0.133 *** 0.016 -0.028 0.175 0.176 ns 0.032 0.077 0.041 0.078
CH04C07 *** 0.005 -0.055 0.064 0.064 *** 0.010 -0.063 0.110 0.108 *** 0.008 0.027 -0.068 0.027
CH01H01 *** 0.037 0.075 0.026 0.028 *** 0.030 0.063 0.034 0.034 * 0.048 0.020 0.083 0.022
Hi02C07 *** 0.003 -0.009 0.102 0.102 ns 0.000 -0.041 0.140 0.140 *** 0.000 0.060 -0.024 0.060
CH01F03b *** 0.028 0.075 0.055 0.055 ns 0.000 0.023 0.084 0.084 *** 0.048 0.022 0.112 0.023
GD147 *** 0.038 0.066 0.102 0.104 *** 0.044 0.073 0.135 0.136 ** 0.025 0.056 0.033 0.058
CH02D08 *** 0.084 0.207 0.114 0.107 ns 0.086 0.203 0.145 0.134 *** 0.085 0.068 0.185 0.063
CH04E05 *** 0.029 0.016 0.455 0.424 ns 0.000 -0.091 0.564 0.564 *** 0.058 0.283 0.010 0.242
CH02C11 *** 0.059 0.131 0.015 0.015 *** 0.065 0.149 0.028 0.026 * 0.048 0.002 0.090 0.004
CH01F02 *** 0.018 0.035 0.035 0.035 ns 0.000 -0.019 0.045 0.045 ns 0.047 0.023 0.101 0.023
CH02C09 *** 0.016 0.036 0.118 0.118 ns 0.000 -0.005 0.137 0.139 *** 0.045 0.079 0.048 0.074
GD12 *** 0.029 0.060 0.062 0.058 *** 0.023 0.029 0.063 0.058 *** 0.035 0.060 0.103 0.056
Mean 0.031 0.059 0.090 0.088 0.023 0.036 0.117 0.116 0.040 0.056 0.063 0.054
a
284 potential M. sylvestris Individuals;
b
167 ‘true type’ M. sylvestris Individuals;
c
117 Hybrids
HWE: ns not significant, * P\0.05; ** P\0.01; *** P\0.001
F
is
and F
st
: bold =significant at P\0.01
Genet Resour Crop Evol
123
Discussion
One fundamental prerequisite for an effective pres-
ervation of the only indigenous wild apple in Europe
is the estimation of population genetic diversity
and the clear identification of hybridization between
M. sylvestris and M. 9domestica. Numerous studies
dealt with this task in which either morphological,
genetic or both parameters were used to differentiate
M. sylvestris from M. 9domestica. In recent years
most of the studies are predominantly based on
molecular markers. If morphological characters are
evaluated they mainly act as a makeshift for com-
parison and assistance purposes in species classifica-
tion (Coart et al. 2003; Koopman et al. 2007; Larsen
et al. 2006).
In this paper the identification of hybrids, evalua-
tion of the genetic variation and differentiation within
and between M. sylvestris and M. 9domestica was
performed based on the statistical analysis of micro-
satellite data.
Admixture analysis and identification of hybrids
A model based analysis including 284 putative
M. sylvestris individuals from the East Ore Mountains,
18 apple cultivars and two further wild apple species
(‘M. floribunda 821’ and ‘M. robusta 5’) were
performed based on 12 SSR marker data in order to
identify M. sylvestris hybrids using the software
STUCTURE. The results show a clear clustering of
all samples in either the ‘M. sylvestris cluster’ or
‘cultivar cluster’ indicating that the M. sylvestris
genotypes are genetically different from M. 9domes-
tica. However, within ‘M. sylvestris cluster’ numerous
individuals were assigned as putative hybrids. This
corresponds to results to other M. sylvestris studies
(Koopman et al. 2007; Coart et al. 2003). In the present
study the defined admixture coefficient threshold of
Q[0.80 to distinguish ‘true type’ M. sylvestris
individuals from ‘hybrids’ was slightly smaller com-
pared to other studies but showed a higher correspon-
dence to the morphological characteristics of the
sampled M. sylvestris individuals. The main part of
M. sylvestris samples were identified as ‘true type’
M. sylvestris (59 %). However, 41 % were assigned as
hybrids and from these 6 % might be feral growing
apple cultivars (Fig. 2).
The proportion of hybrids in the present study is
relatively high compared to other studies (Larsen et al.
2006; Coart et al. 2003). That was expected due to the
fact that putative hybrids were also selected to
estimate the importance of the morphological traits.
Comparison of morphological and genetic traits
The classification of M. sylvestris samples based on
molecular evidence was consistent with the morpho-
logically inferred hybrids or ‘true type’ individuals in
most cases. This was confirmed due to a significant
correlation of the genetic and morphologic distance
matrix calculated by using Mantel test. But, approx-
imately 20 % of M. sylvestris samples showed diver-
gences at the individual level by the classification
based on morphological and genetic data, respec-
tively. In general, the great variation of morphological
characters and sometimes lack of conformity with
molecular data is a known problem (Rieseberg and
Ellstrand 1993; Watano et al. 2004; Larsen et al. 2006;
Coart et al. 2003). Deviations in classifying M. sylves-
tris were also observed in other studies in which the
proportion of morphologically inferred hybrids was
Table 4 Spearman correlation between single morphological
traits and the admixture coefficient (Q) based on genetic data
estimated by STRUCTURE
Morphological trait Admixture
coefficient (Q)
Fruit ground color 0.03
Fruit over color 20.29
Fruit shape -0.17
Depth of stalk cavity -0.14
Width of stalk cavity -0.17
Depth of calyx -0.16
Fruit rust -0.07
Hairiness leaf lower surface -0.01
Hairiness leaf stalk 20.29
Hairiness monopodial shoot 20.35
Hairiness flower stalk 20.44
Hairiness ovary basal 20.44
Hairiness ovary apical 20.42
Hairiness sepal 20.40
Thickness ovary 20.32
Fruit length 20.32
Fruit width 20.34
bold: significant at P=\0.0001
Genet Resour Crop Evol
123
higher than the number of individuals defined as
hybrids by molecular data (Coart et al. 2003; Larsen
et al. 2006). In our study the observation was converse;
the proportion of ‘true type’ M. sylvestris deduced
from the morphological data was higher than from the
molecular data. This fact may be attributed to the high
number of morphological characters that we used for
the evaluation of M. sylvestris in the East Ore
Mountain. In comparison to other studies evaluating
fewer but more descriptive morphological traits,
deviations at single traits had possibly more influence
on the classification of M. sylvestris. In our study, a
precise assessment of the morphological traits in all
individuals showed that the consideration of only one
single character could not be the decisive factor for the
discrimination between ‘true type’ M. sylvestris and
hybrids. However, morphological characters are very
usefull for a first identification of the Malus species
whereas the combination of morphological and
molecular data is recommend for a detailed differen-
tiation of ‘true type’ M. sylvestris and hybrids. In this
context it became clear that the traits ‘flower hairi-
ness’, ‘hairiness on monopodial shoot’ and ‘fruit size’
were most correlated with the genetic inferred ‘true
type’ M. sylvestris individuals and hybrids. The
feature ‘flower hairiness’ seems to be the best
descriptive trait for the classification of M. sylvestris
because it shows the least variation during growing
season.
Genetic variation and differentiation of wild
and cultivated apple
A high level of genetic diversity was found in the ‘true
type’ M. sylvestris population with an expected
heterozygosity of H
e
=0.71. However, this value is
still higher in the apple cultivars (H
e
=0.81) despite
the low number of analysed individuals. The higher
level of heterozygosity of the cultivars is in concor-
dance with former studies on M. sylvestris and
M. 9domestica and may result from the ascertain-
ment bias of the SSRs which were isolated from apple
cultivars (Coart et al. 2003; Larsen et al. 2006;
Koopman et al. 2007).
The near fixation of alleles in the M. sylvestris
population at some SSR loci are particularly reflected
in increased F
st
-values (0.110, 0.137 and 0.564 for
CH04C07, CH02C09 and CH04E05, respectively)
indicating high levels of differentiation between wild
and cultivated apple due to high degrees of fixation on
different alleles. Therefore, the loci CH04C07,
CH02C09, CH04E05 of the ECPGR Malus finger-
printing set are less suitable for the estimation of
genetic diversity in wild apple. Overall, the estimation
of the extent of inbreeding within the complete
M. sylvestris collection, the hybrids and ‘true type’
M. sylvestris population show a slight excess of
homozygotes (F
is
=0.059; F
is
=0.063; F
is
=0.036).
But, the inbreeding coefficient may also be affected by
null alleles (Chapuis and Estoup 2007;Chybickiand
Burczyk 2009; Van Oosterhout et al. 2006). This seems
probably the reason for the high F
is
values within the
complete M.sylvestris collection and the hybrids since
within both groups for each SSR loci the existence of
null alleles has been calculated. Within the ‘true type’
M. sylvestris population for six out of 12 SSR loci show
indications of null alleles (Table 4) which might also
partly explain the slight deficiency of heterozygotes.
Another plausible reason for a positive value of the
inbreeding coefficient is self-fertilization or mating
between related individuals within the M. sylvestris
population. Self-fertilization is usually inhibited due
to the mechanisms of a gametophytic self-incompat-
ibility system (Broothaerts et al. 2004). However, in
small and spatially isolated populations this mecha-
nism can be suspended causing an increased selfing
rate of outcrossing species (De Nettancourt 2001;
Hoebee et al. 2007). In pollination studies small
proportions of self-pollination for apple cultivars
were determined if the apple flowers were isolated by
pollen bags (Sharma and Bashir 2007; Reim et al.
2006). Furthermore, matings between related indi-
viduals may increase with decreasing population
sizes (Ellstrand and Elam 1993) and it cannot be
excluded that a reduction of heterozygosity is caused
by a division of the wild apple collection in several
subpopulations, reproductively isolated by distance
(Wahlund effect).
The population differentiation between the ‘true
type’ M. sylvestris population and the apple cultivars
was statistically significant (with F
st
=0.117) and
remarkable higher after excluding the putative
‘hybrid’ individuals. This result indicated a clear
tendency of genetic fixation of the M. sylvestris
population and the apple cultivars on different allelic
types (Wright 1978). Additionally, a second fixation
Genet Resour Crop Evol
123
index excluding the null alleles was calculated,
because the population differentiation may overesti-
mate in the case of null alleles. This F
st
value
(F
st(FREENA)
=0.116) showed a similar value indi-
cating a low impact of null alleles on differentiation
parameters. Hence, the genetic distance between the
‘true type’ M. sylvestris and M. 9domestica indi-
viduals is in agreement with the separation calcu-
lated by STRUCTURE in this study. The value of
the genetic differentiation parameters of Gregorius
and Nei (D
s
=0.572) also confirms the F
st
and
showed that both species can be differentiated. Other
studies represented similar results regarding the
parameters of differentiation between wild apple
and cultivars (Coart et al. 2003, Larsen et al. 2006).
The genetic differentiation of both species depends
particularly on the extent of hybridization. Despite
of missing crossing barriers, hybridization between
the wild and cultivated apple seems to be restricted.
The principal causes for a low reproductive con-
nectedness are isolation of wild apple trees by
distance and differences in flowering phenology
(Larsen et al. 2006). Generally, it was observed that
M. sylvestris trees flower 10–14 day earlier than
M. 9domestica with some overlap depending on
different years.
The clear genetic distinction between the ‘true
type’ M. sylvestris and apple cultivars indicates that
numerous M. sylvestris individuals were not affected
by gene flow in past times, because of the late
settlement of the East Ore Mountains at the beginning
of the twentieth century. But, despite of the still
existent ‘true type’ M. sylvestris individuals, our study
confirm that hybridization between both species is
taking place endangering the survival of the European
wild apple. Particularly for that reason, it is necessary
to implement conservation strategies in order to ensure
a long-term existence of M.sylvestris in situ. The
occurrence of ‘true type’ M. sylvestris individuals and
the extensive farming in the East Ore Mountains
provide ideal growing conditions for the crab apple
and argue for the implementation of preservation
measures in this area.
Acknowledgments This research project was financial
supported by the Federal Agency for Agriculture and Food for
(06BM 002/2). We thank very much all volunteers from the
Green League East Ore Mountains e.V. for their extensive work
on the Malus sylvestris trees in the East Ore Mountains. We also
thank Claudia Wiedow for her helpful comments.
References
Allendorf FW, Leary RF, Spruell P, Wenburg JK (2001) The
problems with hybrids: setting conservation guidelines.
Trends Ecol Evol 16:613–622
Broothaerts W (2003) New findings in apple S-genotype anal-
ysis resolve previous confusion and request the re-num-
bering of some S-alleles. Theor Appl Genet 106:703–714
Castiglione S, Cicatelli A, Lupi R, Patrignani G, Fossati T,
Brundu G, Sabatti M, van Loo M, Lexer C (2010) Genetic
structure and introgression in riparian populations of
Populus alba L. Plant Biosyst 144:656–668
Cavanna M, Marinoni DT, Bounous G, Botta R (2008) Genetic
diversity in ancient apple germplasm from northwest Italy.
J Hortic Sci Biotech 83:549–554
Chapuis MP, Estoup A (2007) Microsatellite null alleles and
estimation of population differentiation. Mol Biol Evol
24(3):621–631
Chybicki I, Burczyk J (2009) Simultaneous estimation of null
alleles and inbreeding coefficients. J Hered 100:106–113
Coart E, Vekemans X, Smulders MJM, Wagner I, Van Hu-
ylenbroeck J, Van Bockstaele E, Roldan-Ruiz I (2003)
Genetic variation in the endangered wild apple (Malus
sylvestris (L.) Mill.) in Belgium as revealed by amplified
fragment length polymorphism and microsatellite markers.
Mol Ecol 12:845–857
De Nettancourt D (2001) Incompatibility and incongruity. In:
Wild and cultivated plants. Springer, Berlin
Degen B (2008) GDA_NT 2.0: Genetic data analysis and
numerical tests. Available from: bernd.degen@vti.bund.de
Dreesen RSG, Vanholme BTM, Luyten K, Van Wynsberghe L,
Fazio G, Roldan-Ruiz I, Keulemans J (2010) Analysis of
Malus S-RNase gene diversity based on a comparative
study of old and modern apple cultivars and European wild
apple. Mol Breed 26:693–709
Ellstrand NC, Elam DR (1993) Population genetic conse-
quences of small population-size—implications for plant
conservation. Annu Rev Ecol Syst 24:217–242
Evans KM, Fernandez-Fernandez F, Laurens F, Feugey L, van
de Weg WE (2007) Harmonising fingerprinting protocols
to allow comparisons between germplasm collections—
Malus/Pyrus. XIIth EUCARPIA Symposium on Fruit
Breeding and Genetics. Zaragoza (Espagne) 16–20 Sep-
tember 2007 (poster)
Fellenberg U (2001) Beurteilung von Wildobst-Voraussetzung
fu
¨r geeignetes Vermehrungsgut zur Erhaltung von Wald-
genressourcen. Forst und Holz 56:50–54
Gharghani A, Zamani Z, Talaie A, Oraguzie NC, Fatahi R, Ha-
jnajari H, Wiedow C, Gardiner SE (2009) Genetic identity and
relationships of Iranian apple (Malus 9domestica Borkh.)
cultivars and landraces, wild Malus species and representative
old apple cultivars based on simple sequence repeat (SSR)
marker analysis. Genet Resour Crop Ev 56:829–842
Gregorius HR (1974) Genetischer Abstand zwischen Popula-
tionen. I. Zur Konzeption der genetischen Abstandsmes-
sung. Silvae Genet 23:22–27
Guarino C, Santoro S, De Simone L, Lain O, Cipriani G, Testolin
R (2006) Genetic diversity in a collection of ancient cultivars
of apple (Malus 9domestica Borkh.) as revealed by SSR-
based fingerprinting. J Hortic Sci Biotech 81: 39–44
Genet Resour Crop Evol
123
Hedrick PW (2005) Genetics of populations. Jones and Bartlett
Publishers, Sudbury
Hoebee SE, Arnold U, Duggelin C, Gugerli F, Brodbeck S,
Rotach P, Holderegger R (2007) Mating patterns and
contemporary gene flow by pollen in a large continuous
and a small isolated population of the scattered forest tree
Sorbus torminalis. Heredity 99:47–55
Hokanson SC, Szewc-McFadden AK, Lamboy WF, McFerson
JR (1998) Microsatellite (SSR) markers reveal genetic
identities, genetic diversity and relationships in a Ma-
lus 9domestica Borkh. core subset collection. Theor Appl
Genet 97(5–6):671–683
Ho
¨ltken AM, Ta
¨htinen J, Pappinen A (2003) Effects of dis-
continuous marginal habitats on the genetic structure of
common ash. Silvae Genet 52:206–212
Kleinschmit J (1998) Erhaltung und Nutzung wertvoller Edel-
laubbaumarten. Forst und Holz 17:515–519
Kleinschmit J, Stephan R (1997) Wild fruit trees. EUFORGEN
Noble Hardwoods, Network, Reports, pp 51–59
KoopmanWJM, Li YH, Coart E, De Weg EV, VosmanB, Roldan-
Ruiz I, Smulders MJM (2007) Linked vs. unlinked markers:
multilocus microsatellite haplotype-sharing as a tool to
estimate gene flow and introgression. Mol Ecol 16:243–256
Larsen AS, Kjaer ED (2009) Pollen mediated gene flow in a
native population of Malus sylvestris and its implications
for contemporary gene conservation management. Conserv
Genet 10:1637–1646
Larsen A, Asmussen C, Coart E, Olrik D, Kjær E (2006)
Hybridization and genetic variation in Danish populations
of European crab apple. J Tree Genet Genomes 2:86–97
Larsen AS, Jensen M, Kjaer ED (2008) Crossability between
wild (Malus sylvestris) and cultivated (M. 9domestica)
apples. Silvae Genet 57:127–130
Liebhard R, Gianfranceschi L, Koller B, Ryder CD, Tarchini R,
van de Weg E, Gessler C (2002) Development and char-
acterization of 140 new microsatellites in apple (Ma-
lus 9domestica Borkh.). Mol Breed 10(4):217–241
Natzke E, Pech M (1998) Die seltenen Baumarten Wildbirne
und Wildapfel in Sachsen-Anhalt. In: Kleinschmit J, Soppa
B, Fellenberg U. Die Wildbirne, Pyrus pyraster (L.)
Burgsd. Schriften der Forstlichen Fakulta
¨t der Universita
¨t
Go
¨ttingen und der Niedersa
¨chsischen Forstlichen Ver-
suchsanstalt 125: 113–114
Nei M (1972) Genetic distance between populations. Am Nat
106:p283
Nei M (1973) Analysis of gene diversity in subdivided popu-
lations. P Natl A Sci 70:3321–3323
Patzak P (2003) Fo
¨rderung von Wildobst und Feldulme. Beitrag
zum Erhalt der Artenvielfalt der Auewa
¨lder im Biospha
¨-
renreservat Flusslandschaft Mittlere Elbe. Naturwissen-
schaftliche Beitra
¨ge, Museum Dessau 15: 21–43
Paul M, Hinrichs T, Janssen A, Schmitt HP, Stephan BR, Do
¨rf-
linge H (2000) Concept for the conservation and sustainable
utilizationof forest genetic resources in the Federal Republic
of Germany. Sa
¨chsische Landesanstalt fu
¨rForsten,Pirna
Peakall R, Smouse PE (2006) GENALEX 6: genetic analysis in
Excel. Population genetic software for teaching and
research. Mol Ecol Notes 6:288–295
Pritchard JK, Stephens M, Donnelly P (2000) Inference of
population structure using multilocus genotype data.
Genetics 155:945–959
Reim S, Flachowsky H, Michael M, Hanke MV (2006)
Assessing gene flow in apple using a descendant of Malus
sieversii var. sieversii f. niedzwetzkyana as an identifier for
pollen dispersal. Environ Biosafety Res 5:89–104
Reim S, Proft A, Heinz S, Ho
¨fer M (2011) Diversity of the
European indigenous wild apple Malus sylvestris (L.) Mill.
in the East Ore Mountains, Germany: I. Morphological
characterization. Genet Resour Crop Evol. doi:10.1007/
s10722-011-9746-x
Remmy K, Gruber F (1993) Untersuchung zur Verbreitung und
Morphologie des Wild-Apfels (Malus sylvestris (L.) Mill.).
Mitt Deut Dendr Gesells. 81:71–94
Rhymer JM, Simberloff D (1996) Extinction by hybridization
and introgression. Annu Rev Ecol Syst 27:83–109
Rieseberg LH, Ellstrand NC (1993) What can molecular and
morphological markers tell us about plant hybridization.
Crit Rev Plant Sci 12:213–241
Robinson J, Harris SA, Juniper BE (2001) Taxonomy of the genus
Malus Mill. (Rosaceae) with emphasis on the cultivated
apple, Malus domestica Borkh. Plant Syst Evol 226:35–58
Rosenthal G (2003) Bedeutung evolutionsbiologischer Prozesse
fu
¨r Landschaftsplanung und Naturschutz. Natur und Lands-
chaft 78:497–506
Sharma G, Bashir R (2007) Effect of different mode of polli-
nation on fruit set and retention in apple (Malus xdomes-
tica) cultivars. Indian J Agr Sci 77:509–511
Silfverberg-Dilworth E, Matasci CL, Van de Weg WE, Van
Kaauwen MPW, Walser M, Kodde LP, Soglio V, Gianf-
ranceschi L, Durel CE, Costa F, Yamamoto T, Koller B,
Gessler C, Patocchi A (2006) Microsatellite markers
spanning the apple (Malus 9domestica Borkh) genome.
Tree Genet Genomes 2(4):202–224
Slotta TAB, Brady L, Chao S (2008) High throughput tissue
preparation for large-scale genotyping experiments. Mol
Ecol Resour 8:83–87
Spearman C (1904) The proof and measurement of association
between two rings. Am J Psychol 15:72–101
Van Oosterhout C, Weetman D, Hutchinson WF (2006) Esti-
mation and adjustment of microsatellite null alleles in
nonequilibrium populations. Mol Ecol Notes 6:255–256
Wagner I (1996) Zusammenstellung morphologischer Merk-
male und ihrer Auspra
¨gung zur Unterscheidung von Wild-
und Kulturformen des Apfels (Malus) und des Birnbaumes
(Pyrus). Mitt Deut Dendr Gesells. 82:87–108
Wagner I (2006) Malus sylvestris. Enzyklopa
¨die der Hol-
zgewa
¨chse 42:1–16
Watano Y, Kanai I, Tani N (2004) Genetic structure of hybrid
zones between Pinus pumila and P. parviflora var. Pen-
taphylla (Pinaceae) revealed by molecular hybrid index
analysis. Am J Bot 91:65–72
Wright S (1951) Genetical structure of populations. Nature
166:247–249
Wright S (1965) The interpretation of population structure by
F-statistics with special regard to systems of mating.
Evolution 19(3):395–420
Wright S (1978) Evolution and the genetics of populations.
University of Chicago Press, Chicago
Yeh F, Yang R, Boyle T, Ye Z, Xian M (2000) Pogene version
1.31. University of Alberta, Edmonton
Genet Resour Crop Evol
123
