South African Journal of Botany 108 (2017) 175–183

Contents lists available at ScienceDirect

South African Journal of Botany

journal homepage: www.elsevier.com/locate/sajb

Assessment of genetic diversity among six populations of Rhododendron

triflorum in Tibet using ISSR and AFLP markers
J.J. Xu a, L.Y. Zhang b, B. Zhao a,⁎, H.F. Shen a
a
College of Landscape Architecture and Arts, Northwest A&F University, Yangling 712100, China
b
Agricultural and Animal Husbandry College, Tibet University, Linzhi 850400, China

a r t i c l e i n f o a b s t r a c t

Article history: In order to assess genetic variations among Rhodendron trifolium populations sampled from Tibet and determine
Received 12 July 2016 the correlation among genetic variations, the geographic location of a population, and factors that influence
Received in revised form 6 September 2016 high-level genetic diversity, in total of 107 R. triflorum samples using inter simple sequence repeats (ISSR) and
Accepted 13 October 2016
amplified fragment length polymorphisms (AFLP). All genotypes were collected from six different areas. Eleven
Available online 28 October 2016
ISSR primers and five AFLP primer pairs were screened, and 118 and 169 amplification products were produced,
Edited by J Van Staden of which, 96.61% in ISSR and 95.27% in AFLP were polymorphic. High genetic diversity was observed at the species
level: Nei's genetic diversity (H) was 0.3382 and 0.306, and Shannon's information index (I) was 0.5085 and
Keywords: 0.4642 in ISSR and AFLP, respectively. Both the coefficient of gene differentiation (GST 0.3752 in ISSR, 0.31 in
Rhododendron triflorum AFLP) and AMOVA analysis (75% in ISSR, 71% in AFLP) indicated that most genetic diversity was distributed
Genetic diversity within populations. Gene flow (Nm) was 0.8326 in ISSR and 1.1127 in AFLP. The analyze of unbiased genetic
Habitat types distances determined by an unweighted pair group method using arithmetic mean (UPGMA) phonograms
Gene flow indicated that there was a certain degree correlation between the genetic distance and the geographic distance,
Conservation strategies
which was confirmed by a principal coordinate analysis. The results maybe indicated that difference of geograph-
ical environment and variation of habitat types caused the genetic differentiation of R. triflorum. At last, some
conservation strategies for R. triflorum germplasm were put forward according to these results.
© 2016 SAAB. Published by Elsevier B.V. All rights reserved.

1. Introduction to use population genetics of individual species to establish strategies

The Tibetan and adjacent regions of China was one of the most
diverse alpine plants region in the world (Yu and Zhang, 2013). Ti-
betan provides the best habitat for growing azaleas and, together
with Yunnan and Sichuan (Geng, 2010), forms the azalea distribu-
tion center of the world. It is important to understand the resources
of Rhododendron in Tibetan. Rhodendron trifolium, the species in this
study, is a perennial evergreen dwarf shrub inhabiting alpine re-
gions, and has a high ornamental value. As R. triflorum is a hardy or-
namental plant that grows in the wild and has the potential to enrich
the alpine plant community and increase plant material for use in
constructing high altitude landscapes, both Xiao et al. (2010) and
Xing et al. (2011) suggested that this plant should be preferentially
developed. Guidelines of reasonable exploitation and utilization of
the germplasm resources should have been established to protect
the germplasm resources of R. triflorum.
Determination of genetic diversity is essential for conservation and
increased exploitation of genetic resources (Dvorakova et al., 2015).
Genetic diversity differs widely among species; therefore, it is necessary
⁎ Corresponding author.
E-mail address: bingbing2003915@163.com (B. Zhao).
for reasonable utilization and protection (Zhao et al., 2012). In addition,
determining genetic diversity can help identify the genotypes and
horticultural characteristics of the species, and can help to establish
methods for gene transfer, which can shorten breeding programs
(Okcu et al., 2015). Presently, research on genetic diversity is an impor-
tant means for protecting and utilizing germplasm, and it provides an
important basis for breeding. Liu et al. (2012) demonstrated that life
forms, taxonomic status, and geographic locations of plants strongly
impact genetic variation, and that these factors partition the genetic
diversity within and among plant populations. In studies in genetic
diversity of a wild species of sorghum, Basahi (2015) demonstrated
that genetic diversity is influenced by human and natural factors,
including altitude, soil, climate, and gene flow. Different altitudes and
regions may restrict gene flow between the populations in different
geographical locations, and then may lead to genetic differentiation
(Arnaud-Haond et al., 2006). Recently, many studies of genetic diversity
have focused on geographic variation, such as elevation gradients
(Thiel-Egenter et al., 2009; Liu et al., 2012). Assessment of the genetic
diversity in populations that belong to different geographic regions
could reveal gene flow and indicate strategies for utilization and
conservation.
A number of molecular markers have been used successfully to
evaluate genetic relationship (Abraha et al., 2016) and genetic diversity

http://dx.doi.org/10.1016/j.sajb.2016.10.023
0254-6299/© 2016 SAAB. Published by Elsevier B.V. All rights reserved.
176 J.J. Xu et al. / South African Journal of Botany 108 (2017) 175–183
of natural plants (Mokhtari et al., 2013; Brandao et al., 2015; Larsen
et al., 2015; Moreira et al., 2015; Okcu et al., 2015; Basu et al., 2016;
Ngailo et al., 2016). Additionally, many molecular markers had been
used to study the genetic diversity of wild Rhododendron resources,
including horizontal starch gel electrophoresis, RAPD, ISSR (Liu et al.,
2012), SSR (Li et al., 2015), and AFLP (Tikhonova et al., 2012). Therefore,
it is necessary to choose the most suitable technique to assess the genetic
diversity of R. triflorum, based on cost and qualities of the techniques.
ISSR is a molecular markers technique based on inter-tandem
repeats of short DNA sequences (Al-Turki and Basahi, 2015), and is
widely used to determine genetic distances between organisms. ISSR
has many advantages, including simplicity, reproducibility, and low
cost. It does not require information about the DNA sequences
(Morshedloo et al., 2015), and it can produce a large number of frag-
ments per primer. ISSR uses the SSR motifs anchored with 2–4 random
nucleotides as primers, which are anchored to genomic sequences to
generate either side of the targeted simple sequence repeats (Al-Turki
and Basahi, 2015). Compared to RAPD, ISSR can detect a greater level
of genetic diversity (Dvorakova et al., 2015). Wolfe and Liston (1998)
demonstrated that ISSR can provide more reliable and reproducible
results than RAPD when using longer primers and higher annealing
temperatures.
Powell et al. (1996) demonstrated that AFLP is a suitable technology
for assessing genome-wide marker distributions and estimating genetic
diversity. Vos et al. (1995) also demonstrated that AFLP is a powerful
method for detecting DNA polymorphisms and marker analysis,
because the primer combinations allow analysis of a large number of
loci and have high polymorphism as the same as reproducibility. The
primary step AFLP is the digestion step, in which the DNA is digested
into fragments by restriction enzymes without requiring sequence
information. Then, the fragments are selected and amplified. Zhang
et al. (2014) proposed that AFLP markers are currently the better choice,
if sequence information is not known.
To our knowledge, research reports of the genetic diversity of
R. triflorum have never been published. Moreover, although ISSR and
AFLP markers have never been used to study R. triflorum, they have
been used to study other species, including Rhododendron aureum
Georgi (Liu et al., 2012), Rhododendron species from Far East of Russia
(Tikhonova et al., 2012), and Rhododendron ferrugineum (Escaravage
et al., 1998). ISSR and AFLP are promising technologies for the estima-
tion of genetic diversity with high degrees of reliability (Zhao et al.,
2012; Jena et al., 2015; Kumar et al., 2015). Furthermore, there have
been assessments of genetic diversity in other plants, based on ISSR
and AFLP markers (Das et al., 2015).
Keeping into perspective view about the above discussed facts, the
present study was to clarify the genetic diversity of R. triflorum using
ISSR and AFLP markers, as there is not any information of the genome
of this species. Moreover, the efficiency and usefulness of these two
techniques in estimating genetic variation were compared. Ultimately,
the results of these two techniques were compared and complemented
mutually, in order to evaluate the genetic diversity of R. triflorum more
comprehensively. This analysis is expected to provide the information
needed to improve future methods for germplasm collection and to
help establish reasonable strategies for the utilization, management,
and conservation of R. triflorum.
2. Materials and methods
2.1. Plant materials
Plant materials of R. triflorum used in the present study totaled 107
and were collected from 6 locations of Tibet which were coded as
Milin1 (M), Milin2 (ML), Bomi (B), Sejila Mountain (S), Linzhi (LH),
and Galongla (G). The detailed information of 6 populations is enumer-
ated in Table 1. Cluster sampling method has been used to sample the
materials; samples were always at least 3 m away from each other to
Table 1
Overview of Rhododendron triflorum populations, located in different regions of Tibetan
China.
No. Population Sample size Coordinates Altitude(m)
1 Milin1 (M) 19 E94°00′48″ N29°11′22″ 2984
2 Milin2 (ML) 20 E94°16′33″ N29°14′56″ 2965
3 Bomi (B) 17 E95°21′47″ N29°57′52″ 2610
4 Galongla (G) 20 E95°34′39″ N29°35′06″ 3543
5 Sejila (S) 15 E94°43′36″ N29°49′27″ 3526
6 Linzhi (LH) 16 E94°19′36″ N29°40′17″ 2980
avoid sampling ramets from the same vegetative clone (Liu et al.,
2012). Young leaves were sampled from plants of R. triflorum,
and dried directly in silica gel for transportation, then been frozen at
−20 °C for storage till DNA extraction.
2.2. DNA extraction
Total genomic DNA was extracted from 0.03 g powdered leaf mate-
rials using a plant genomic DNA Kit (Tiangen, Beijing, China) according
to the manufacturer's instructions. The purity and the quality of the
extracted DNA were detected by 1% (w/v) agarose gel electrophoresis
and a UV spectrophotometer (Liuyi, Beijing, China). The gels were
stained with ethidium bromide (EtBr and EB) and photographed
using a Bio-rad gel documentation system (Bio-Rad Laboratories,
Shanghai, China). Finally, each DNA was diluted to 50 ng μL− 1 and
stored at −20 °C for further fingerprinting analysis.
2.3. ISSR amplification
The ISSR analyses were performed according to Liu et al. (2012)
with minor modifications. A total of 40 primers were designed by the
University of British Columbia (UBC) and synthesized by AuGCT (Beijing,
China). 11 primers, which could produce strong, clear and reproducible
bands, were screened from 40 primers using 6 individuals from 6 popula-
tions and used for the amplification reactions for all samples in this study
(Table 2). Briefly, ISSR-PCR amplification reactions were performed
following Xu et al. (2016), This experiment had been repeated three
times.
2.4. AFLP amplification
The AFLP analyses in this study were carried out following Vos et al.
(1995) and Xu et al. (2016) with some adjustment as. Selective amplifi-
cations were performed in 20 μL volume by using 5 selected primer
combines (Table 2) and conducted in a gradient type PCR machine
with the following profile: 4 min at 94 °C; using 12 cycles of 30 s at
94 °C, 30 s at 65 °C (lowering the annealing temperature by 0.7 °C
over each cycle) and 1 min at 72 °C; then 24 cycles of denaturation at
94 °C for 30 s, annealing at 56 °C for 30 s, extension at 72 °C for 1 min;
extension at 72 °C for 8 min at last.
For estimation, the products of selective amplification, which were
mixed with 7 μL denaturation buffer, were denatured for 8 min at
98 °C and cooled on ice immediately. Each 4 μL of the denatured prod-
ucts was separated by 6% (w/v) polyacrylamide gel electrophoresis
(PAGE) in 1 × TBE buffer using 100 bp Ladder (Cwbiotech, Beijing,
China), stained by silver staining solution ultimately.
2.5. Genetic analysis and construction of dendrogram
For the genetic diversity analysis, ISSR and AFLP amplified fragments
which are distinct, reproducible and unambiguous were assembled into
binary matrix by manually scoring present band as (1) or absent band as
(0). The resulting binary matrix was subjected to estimate genetic
parameters using the genetic software program POPGENE version 1.31
(Yeh et al., 1997), including observed number of alleles (Na), effective
 J.J. Xu et al. / South African Journal of Botany 108 (2017) 175–183 177

Table 2
Oligonucleotide sequences, total number of bands and polymorphic bands of ISSR primers
and AFLP primer combinations.

No. Primer sequence 5′-3′ Amplification bands Polymorphism bands PPL%

ISSR
1 UBC808(AG)8C 11 11 100
2 UBC810(GA)8T 13 12 92.31
3 UBC811(GA)8C 8 8 100
4 UBC817(CA)8A 6 6 100
5 UBC823(TC)8C 9 8 88.89
6 UBC827(AC)8G 9 9 100
7 UBC834(AG)8YT 13 13 100
8 UBC836(AG)8YA 10 11 90.91
9 UBC841(GA)8YC 13 12 92.31
10 UBC842(GA)8YG 15 15 100
11 UBC845(CT)8RG 10 10 100
Total 118 114 96.61

AFLP
1 54–44 E-AGC/M-CTG 28 26 92.86
2 52–42 E-ACC/M-CAT 35 33 94.29
3 54–42 E-AGC/M-CAT 39 37 94.87
4 53–46 E-ACG/M-CTT 33 32 96.97
5 48–43 E-AAC/M-CTA 34 33 97.06
Fig. 1. Location of six populations of Rhododendron triflorum included in this study.
Total 169 161 95.27

number of alleles (Ne), number of polymorphic fragments (P), percent
polymorphism Loci (PPL), Nei's genetic diversity (H, Nei, 1973),
Shannon's information index (I, Lewontin, 1972), coefficient of gene
differentiation (GST) and gene flow (Nm, McDermott and McDonald,
1993). Additionally, analysis of molecular variance (AMOVA) was used
to calculate the population genetic variation with and among 6 popula-
tions based on the difference of geographical region by using GENALEX
version 6.5 (Peakall and Smouse, 2012). Ultimately, using NTSYSpc
version 2.10 (Rohlf, 2000), a matrix of Nei's unbiased genetic coefficient
of 6 populations was used to construct a UPGMA dendrogram which
was used to illustrate the relationships among these populations, and
the data was also estimated using a principal coordinate analysis
(PCO) to display the distribution of populations.
3. Results
3.1. Genetic diversity analysis
The heredity of six populations was analyzed using the following
common parameters: effective number of alleles (Ne), number of poly-
morphic fragments (P), percent polymorphism loci (PPL), Nei's measure
of genetic diversity (H), and Shannon's information index (I). In the
present study, 118 scored bands, ranging from 2000 bp to 100 bp,
were amplified using 11 ISSR-selected primers (Table 2, Fig. 1.). The
number of generated bands per primer ranged from 15 (UBC842) to
six (UBC817), and the average was 10.7 bands per primer. Of the
bands, 114 were polymorphic fragments found among 107 individuals,
which corresponded to 96.61% polymorphism. The number of polymor-
phic bands ranged from 15 (UBC842) to six (UBC817), with an average
of 10.3 bands per primer. The PPL of all individuals ranged from 100% to
88.89% and the PPL of each population ranged from 63.56% to 82.2%,
with an average of 72.60%. With suitable E/M primer combinations
selected, AFLP was performed successfully for R. triflorum. Out of 64
primers pairs, five were used to selectively amplify 107 fragments
(Table 2, Fig. 2.). Whereas the five primer pairs, 169 bands, ranging
from 1000 bp to 100 bp, were amplified, of which, 161 were polymor-
phic. Based on banding pattern, the AFLP primer pair, E-AGC/M-CAT
generated the highest number (39 fragments) of bands and the primer
pair E-AGC/M-CTG generated the lowest number (28 fragments), with
an average of 33.8 fragments per primer pair. The PPL of the species
based on different primer combinations ranged from 97.06% (E-AAC/
M-CTA) to 92.86% (E-AGC/M-CTG) and the PPL of each population
ranged from 86.39%, for G, to 52.07%, for LH, with an average of
69.92%. The order from the highest to lowest PPL based on ISSR was as
follows: M N B N G N ML N S N LH, but the order based on AFLP was G N
M N ML N B N S N LH. The value for Nei's measure of genetic diversity
(H) for the six populations was 0.3382 based on ISSR and 0.3060
based on AFLP, and Shannon's information index (I) of the species was
0.5085 and 0.4642, based on ISSR and AFLP, respectively. Based on the
results from the two techniques, M exhibited the highest level of diver-
sity with the H = 0.2645 (ISSR), H = 0.2467 (AFLP), I = 0.4008 (ISSR)
and I = 0.3761 (AFLP). In addition, both Shannon's information index
(I) and Nei's measure of genetic diversity (H) of M were the highest in
ISSR followed by those in AFLP. The orders of I of the population, from
the highest to lowest, were as follows: M N ML N B N G N S N LH (ISSR)
and G N M N ML N S N LH N B (AFLP), and the orders of H of population,
from the highest to lowest, were M N ML N G N B N S N LH (ISSR) and G N
M N ML N S N LH N B (AFLP). In the calculation of ISSR, the two indicators
of B were almost equal to those of G. Thus it can be postulated that the
genetic variations in the six R. triflorum populations determined by
Shannon's information index (I) were consistent with those determined
by Nei's measure of genetic diversity (H), either in ISSR or AFLP; however,
the order of genetic diversity indicated some differences between ISSR
and AFLP.
3.2. Cluster analysis
To analyze the genetic differentiation among the populations further,
a UPGMA dendrogram of genetic relationship was constructed based on
the Nei–Li coefficient matrices of the ISSR and AFLP (Figs. 3, 4.). The
genetic distance between every two populations was larger than zero,
which indicated that the six populations shared the same genetic
background. The coefficient range of the different populations was
from 0.9242 to 0.6808, by ISSR, and from 0.9727 to 0.729 by AFLP. In
this dendrogram, all six populations were classified into two main
branches in both ISSR and AFLP. In the dendrogram of ISSR (Fig. 3), B
and G were in one group and the other populations were in the other
group. Moreover, when the coefficient was equal to 0.92, the six popula-
tions were clustered into three groups: one group consisted of M and ML,
one consisted of LH and S, and one consisted of B and G. However, LH and
S were clustered into a group with M and ML in the dendrogram of AFLP
(Fig. 4.). The four populations who shared the same branch were
classified into three groups when the coefficient was equal to 0.96: one
group consisted of G, one consisted of B, and one consisted of M and ML.
The Nei–Li coefficient matrix, based on ISSR and AFLP assessments,
was subjected to principal coordinate analysis (PCO, Fig. 5a and b.).
The distribution of the six populations is shown in three dimensional
178 J.J. Xu et al. / South African Journal of Botany 108 (2017) 175–183
Fig. 2. ISSR pattern of Milin2 population obtained by the primer UBC842.
scatter plots, which provide a spatial representation among the popula-
tions. The information about the spatial distribution of the six popula-
tions was in consensus with the tree plots. The B and G populations
were clearly grouped apart from the others in the PCO plot of ISSR
(Fig. 5a.), and the M and ML were grouped together. The PCO plot of
AFLP differentiated S and LH from the remaining populations, of
which M and ML were also grouped close to each other.
3.3. Genetic differentiation of populations from different geographical areas
As summarized in Table 3, the genetic differentiation coefficient
(GST) for the species, based on the six populations sampled in this
study, was estimated to be 0.3752 by ISSR, indicating that 37.52% of
the genetic variation was distributed among the populations and
62.48% was within the populations. Based on AFLP, the GST of all
populations was 0.31, which indicated that 31% and 69% of the genetic
variation was distributed among populations and within populations,
respectively. In general, genetic variation was high within populations,
whether tested by markers for ISSR or for AFLP. The gene flow (Nm),
calculated as Nm = 0.5 ∗ (1 − GST) / GST, was 0.8326 and 1.1127
when using the ISSR and AFLP markers, respectively. The results from
AMOVA analysis (Table 4) were consistent with Nei's genetic diversity
(Table 3). AMOVA showed that 75% and 71% of variance occurred within
600bp
500bp
400bp
300bp
200bp
100bp
Fig. 3. AFLP pattern Bomi population obtained by the selective primer combination
E-ACC/M-CAT (100–600 bp).
2000bp
1500bp
1000bp
750bp
500bp
250bp
100bp
populations, based on ISSR and AFLP markers respectively, and that 25%
and 29%, of genetic variation, respectively, occurred among populations.
AMOVA also indicated that a high level of genetic variation occurred
within populations. The mean values of the overall FST for individual
loci, estimated by ISSR and AFLP, were 0.254 and 0.291, respectively.
These values are significantly different from zero (P ≤ 0.001).
To analyze genetic diversity further, the six populations were placed
into different groups, correlating with those determined by tree plots,
and then GST and Nm of each group were estimated by POPGENE
(Table 5). Based on ISSR, the genetic differentiation coefficient (GST)
ranged from 0.3662 to 0.0913 and the number of migrants (Nm) ranged
from 4.9976 to 0.8653. Based on AFLP, the GST ranged from 0.2935 to
0.0414, and the Nm ranged from 11.5779 to 1.2034. As summarized in
Table 5, the lowest level of variation based on ISSR was Gr. (M, ML)
and Gr. (S, LH), and the highest numbers of migrants (Nm) were Gr.
(M, ML) and Gr. (S, LH). Based on AFLP, the lowest level of variation
indicated by GST was Gr. (M, ML), which was also the group with
the highest number of migrants (Nm). In general, the order of Nm is
opposite that of GST, and the genetic variation of groups with four
populations was higher than that of groups with two populations.
4. Discussion
In the present study, we analyzed the genetic diversity of six popula-
tions of R. triflorum that are found in the Himalayas and Nyainqêntanglha,
at altitudes ranging from 2610 m to 3543 m. There are previous reports on
the genetic diversity on Rhododendron; however, to the best of our knowl-
edge, research on genetic diversity of R. triflorum has not been reported.
In this study, a number of fragments and a high level of PPL were
amplified both by ISSR and AFLP primers (Table 2). PPL was found to
be 98.31% (ISSR) and 95.86% (AFLP) in species level, and I was found to
be 0.5085 (ISSR) and 0.4642 (AFLP), which were higher than that Liu
et al. (2012) used ISSR to examine the genetic diversity of R. aureum lo-
cated in the Changbai Mountain (PPB = 87.43% , I = 0.4593) and slightly
lower than the values obtained by Zhao et al. (2012) which found that
the genetic diversity of five Rhododendron species (PPL = 92.94%, I =
0.5565). These results indicated that the techniques were used success-
fully to survey the genetic diversity of six populations of R. triflorum. In
species level the Hÿwas found to be 0.3382 (ISSR) and 0.306 (AFLP)
which were higher than the average genetic diversity index (H =
0.2200) statistics by Nybom (2004). It indicated that all the sampled
R. triflorum had high level genetic diversity. In general, differences in
life-history traits likely lead to differences in the mode and speed of evo-
lution between trees and seed-propagated, selfing annuals (Cornille
et al., 2012). Moreover, it was pointed out that life form, mating system,
and environments are pivotal factors that influence genetic diversity of a
species (Liu et al., 2012; Chen et al., 2016; Dulya and Mikryukov, 2016).
R. triflorum is a long-lived, perennial, evergreen, alpine dwarf shrub with
a long juvenile phase and outcrossing and selfing abilities. One character-
istic common to alpine plants is the ability to reproduce both sexually
 J.J. Xu et al. / South African Journal of Botany 108 (2017) 175–183
0.71 0.77 0.83
Coefficient
(a)
0.78 0.83 0.88
Coefficient
(b)
M--Milin1,ML--Milin2, B--Bomi, S--Sejila, LH--Linzhi, G--Galongla.
Fig. 4. UPGMA dendrograms for six Rhododendron triflorum populations based on ISSR (a) and AFLP (b). M—Milin1, ML—Milin2, B—Bomi, S—Sejila, LH—Linzhi, and G—Galongla.
and asexually by sprouting prostrate stems in order to adapt to drastic
environments (Rafii and Dodd, 1998; Liu et al., 2012). According to pre-
vious research, the life form, mating system, and environment including
the influence of human activities, were used to explain the high genetic
diversity in R. triflorum in the present study. First, different regions of
R. triflorum populations from the Himalayas and Nyainqêntanglha have
diverse ecological environments, which lead to variation in the types of
population habitats. The activities of humans have resulted in a signifi-
cant reduction in the resources of most plants, including R. triflorum.
Second, the long juvenile phase of R. triflorum provides increased oppor-
tunities for asexual reproduction. Abundance of natural vectors leads to a
high level of genetic diversity in populations (Hamrick and Godt, 1996;
Zhao et al., 2012). Because of entomophilous pollination and wind
pollination, exchange of these vectors frequently occurs between an
individual, which maintains high genetic variation.
The assessments using POPGENE (Table 3) indicated that the genetic
diversity levels of populations M, ML, and G were higher than those
populations S, LH, and B, which have smaller sample sizes than others.
Previous studies have discussed how population size and sample size
affect genetic diversity and concluded that smaller populations tend to
have lower genetic diversity than larger populations (Frankham,
1996; Godt et al., 1996; Trapnell and Hamrick, 2005; Liu et al., 2012;
Andrello et al., 2016). LH, the smallest of the six populations, was dem-
onstrated with the lowest genetic diversity. The results supported the
aforementioned theory and indicated that population size was a factor
related to genetic variance. Small populations tend to have increased
179
ML
LH
0.90 0.96
ML
LH
0.92 0.97
opportunities for inbreeding and clonal crossing, which probably
reduces gene resources (Ellstrand and Elam, 1993). In general, inbreed-
ing and clonal crossing not only would increase opportunities for the
pairing and expression of recessive alleles that maybe the harmful
alleles for the population's fitness, but also would augment gene drift.
Moreover, the ability of low frequency alleles to enhance the adaptabil-
ity of a population would be lost (Lande et al., 1999; Liu et al., 2012).
Natural selection would reduce the population by eliminating unfitness,
and thereby reduce the genetic variation of the population. Another
factor related to genetic variance was environment. The difference
between LH and other populations was that LH was collected from a
mountain that is located in a town. Human activities in the region
inhabited by LH were higher than those in the locations of other popu-
lations, and these activities would influence the native condition of the
population, even destroying the wild germplasm of population LH for
market transactions. Phenotypes and genotypes that promote survival
were lost as a result of human activities, including collecting the plants
for use as ornaments; consequently, this loss has impeded gene fitness
and reduced the genetic diversity of the population.
To assess population genetic structure, the genetic differentiation
coefficient among populations (GST) was calculated using POPGENE.
Both the ISSR (GST = 0.3752) and AFLP (GST = 0.31) indicated that
the genetic diversity mainly belonged within populations, which was
further confirmed by the AMOVA results. As indicated by AMOVA, 75%
and 71% of variance occurred within populations based on ISSR and
AFLP markers, respectively, and only 25% and 29%, respectively, of the
180 J.J. Xu et al. / South African Journal of Botany 108 (2017) 175–183

(a)

(b)
M--Milin1,ML--Milin2, B--Bomi, S--Sejila, LH--Linzhi, G--Galongla.

Fig. 5. Three-dimensional plot of the principal coordinates analysis of six Rhododendron triflorum populations based on ISSR (a) and AFLP (b). M—Milin1, ML—Milin2, B—Bomi, S—Sejila,
LH—Linzhi, and G—Galongla.

Table 3
Genetic diversity of Rhododendron triflorum populations distributed in different regions, detected by ISSR and AFLP markers.

Population PPL% Na Ne H I Ht Hs GST Nm

ISSR
Milin1 (M) 82.2 1.8221 1.4452 0.2645 0.4008
Milin2 (ML) 68.64 1.6864 1.3667 0.2175 0.3295
Bomi (B) 80.51 1.8051 1.3084 0.2006 0.3211
Galongla (G) 72.03 1.7203 1.3367 0.2069 0.3204
Sejila (S) 68.64 1.6864 1.2867 0.1818 0.287
Linzhi (LH) 63.56 1.6356 1.2820 0.1771 0.2773
Mean 98.31 2.0000 1.5751 0.3382 0.5085 0.3330 0.2081 0.3751 0.8326

AFLP
Milin1 (M) 77.51 1.7751 1.4109 0.2467 0.3761
Milin2 (ML) 73.96 1.7396 1.3879 0.2342 0.3574
Bomi (B) 67.46 1.6746 1.2079 0.138 0.2280
Galongla (G) 86.39 1.8639 1.4098 0.2517 0.3887
Sejila (S) 62.13 1.6213 1.3932 0.2257 0.3345
Linzhi (LH) 52.07 1.5207 1.3170 0.1821 0.2706
Mean 95.86 1.9586 1.5138 0.306 0.4642 0.30 0.2131 0.3099 1.1127

Ne = Effective number of alleles (Kimura and Crow, 1964).

H = Nei's (1973) gene diversity.
I = Shannon's Information index (Lewontin, 1972).
PPL = Percentage of polymorphic loci.
Ht = average genetic diversity within species.
Hs = average genetic diversity within populations.
GST = proportion of species genetic diversity attributed to among-population variation. GST = (Ht _ Hs)/Ht.
Nm = gene flow.
 J.J. Xu et al. / South African Journal of Botany 108 (2017) 175–183 181

Table 4 was located in Nyainqêntanglha, and M and ML were located in a complex

Genetic variation within and among populations of Rhododendron triflorum revealed by region between the Himalayas and Nyainqêntanglha. In short, complex
AMOVA, based on ISSR and AFLP markers.
environments resulted in variation among populations. Thus, the fact
Source of df Sum of MS Variance Percent of P-value that B and G were not clustered into groups with other populations in
variation squares component variation the AFLP-based analysis, unlike the ISSR-based analysis, can be explained
ISSR by extreme differences in altitude.
Among populations 5 517.010 103.402 4.955 25% 0.001 To further analyze the genetic diversity of the populations and to
Within populations 102 1484.749 14.556 14.556 75%
determine the reasons for genetic differentiation among populations, a
Total 107 2001.759 19.512 100%
series of genetic structure data was estimated by POPGENE and
AFLP GENEALEX. As seen in the analysis of genetic structure, the results of
Among Pops 5 903.187 180.637 8.930 29% 0.001
groups based on clustering indicated that the genetic variation within
Within Pops 101 2198.327 21.766 21.766 71%
Total 106 3101.514 30.696 100% a population was dominant. These results supported a previous study
that evaluated the genetic structure of populations from different
altitudes and indicated that genetic differentiation is consistent with
genetic variation among populations. These results are similar to those the mating system of R. aureum (Liu et al., 2012). The previous report
from previous studies (Li et al., 2015; Stout et al., 2015; Wu et al., suggested that the level of effective gene flow with more than four
2015). Geographical isolation and long-term natural selection allowed migrations per generation was enough to offset gene drift, thereby,
the local population to maintain a specific genotype, thereby increasing preventing genetic differentiation between populations (Slatkin, 1987;
the genetic differences between populations (Ng and Corlett, 2000; Li Liu et al., 2012). Assessments of two markers by using POPGENE and
et al., 2015); however, the low frequency of gene exchange, caused by GENEALEX indicated that genetic diversity among population accounts
geographical isolation, lead to a reduction in genetic diversity. Further- for approximately one third, which was much higher than that reported
more, the habitation of a small geographic region by a population makes by previous studies (Zhao et al., 2012). Analysis of the genetic structure
pollination among individuals of that population easier. Chappell et al. of whole populations indicated that the gene flow of R. triflorum was
(2008) suggested that genetic diversity within populations would be 0.8326 (ISSR) and 1.1127 (AFLP), which were similar to the gene flow
higher if vectors were limited to the pollination of only part of the in R. aureum (Liu et al., 2012). These data indicated although a certain
whole population. The populations in the present study grew in a region amount of gene exchange occurred among populations located in differ-
rich with plant species and isolated by a coniferous forest. As a result, ent areas, gene flow was not the only factor leading to high genetic
pollination among the whole population was limited, so that a high diversity. Gene flow between populations in the same group was
level of genetic variation occurred within the population. Additionally, analyzed to confirm the results based on clustering. As indicated by
Wright (1951) suggested that the genetic structure of outcrossing the analysis of groups M and ML, there was little diversity, but signifi-
species would be lower than that of inbreeding species. Thus, the cant gene flow, among populations. The gene flow between B and G
outcrossing reproductive system of R. triflorum may cause low genetic was approximately 3.9307 (ISSR) and 3.5867 (AFLP), and the GST was
variation among populations. 0.1128 (ISSR) and 0.1223 (AFLP), indicating that altitude maybe a lead-
To analyze genetic differentiation further, a dendrogram of genetic ing factor in high genetic diversity and low gene exchange between
relationships was developed, in which all the populations were separated populations. On comparing the order of GST with Nm for each group, it
into two main branches, based on both the ISSR and AFLP results. The was found that increased gene flow was associated with reduced gene
dendrogram from the ISSR results indicated that geographic distance differentiation between populations, which indicated that gene flow
significantly correlated with genetic differences among the six popula- was a significant, but not the only, factor effecting genetic diversity. An-
tions. The six populations first were divided into two main groups, other important factor influencing the distribution of genetic variation
and then were further divided into three groups based on different among populations may be the local environmental conditions.
longitude and latitude. Geographic distance also seemed to significantly The present study on the genetic diversity of R. triflorum was helpful
affect genetic clustering. The results showed that the six populations for developing effective strategies for conserving genetic resources and
were separated into different groups, in which the nearest populations for further breeding. Calculated by ISSR, the genetic diversity of M was
were grouped together, but the populations M and ML were grouped high and the others were low. Whereas the genetic diversity of M, ML,
with the distant populations, B and G. These results can be explained G and S were high and B and LH were low which were revealed by
as follows. Chen et al. (2016) suggested that the geographic distribution AFLP. Zhao et al. (2012) suggested that the genetic diversity within pop-
of species and the representative sample size could be pivotal factors in ulations should be considered for germplasm conservation and that
genetic diversity. Fewer differences in climate and environment stress populations with high diversity should be chosen first for in situ conser-
caused by shorter geographic distance might cause less variation in vation because they can maintain the greatest degree of genetic diversi-
the population, including phenotype and genotype. However, the ty. Furthermore, global climate change and the impact of the activities of
sampling region for this study was isolated by the Himalayas and the human beings are increasingly influencing the evolution and conserva-
Nyainqêntanglha Mountains, so that the climate and environment was tion of germplasm, especially in small plant populations. As proposed in
quietly diverse. B and G were located on the windward side of Himalayas, a previous study, mature individuals that maintain genetic diversity and
S was located on the eastern slope of the Sejila Mountain, which, like LH ecological adaption in populations should be first conserved to protect
reproductive fitness and evolutionary potential of those populations
(Cruse-Sanders et al., 2005, Wu et al., 2015). Because each population
Table 5 had unique genotypes and because gene flow was significant in main-
Genetic diversity of Rhododendron triflorum populations, belonging to the same group
taining heterosis of Rhododendron populations, populations from differ-
detected by ISSR and AFLP markers.
ent geographic locations should be collected for germplasm and ex/in
Ht Hs GST Nm Ht Hs GST Nm situ conservation in order to obtain healthy germplasm and to preserve
ISSR AFLP genetic diversity (Lienert, 2004, Devey et al., 2009 and Wu et al., 2015).
M, ML 0.2700 0.2410 0.1074 4.1552 0.2508 0.2405 0.0411 11.6748 Considering the results from the analyses of genetic diversity, genetic
S, LH 0.1975 0.1795 0.0911 4.9861 0.2422 0.2039 0.1581 2.6619 structure and the population size, population M should be chosen for
B, G 0.2296 0.2037 0.1128 3.9324 0.2221 0.1949 0.1225 3.5827 in situ conservation at first. Other populations, LH and B especially,
M, ML, B, G 0.3509 0.2224 0.3662 0.8654 0.2497 0.2177 0.1282 3.4016 should be chosen for in situ conservation and ex situ conservation in
M, ML, S, LH 0.2578 0.2102 0.1846 2.2080 0.3145 0.222 0.2941 1.2000
order to improve the level of genetic diversity and to prevent species
182 J.J. Xu et al. / South African Journal of Botany 108 (2017) 175–183
extinction. In addition, the assessment of genetic diversity of R. triflorum
should be strengthened and domesticated and cultivated R. triflorum
should be enhanced, which could lay the foundation of the protection
and utilization of R. triflorum.
5. Conclusions
Using both ISSR and AFLP markers, the present study examined the
genetic diversity and structure of R. triflorum populations located in
different regions of Tibet successfully. The results indicated a high
level of genetic diversity at the species level, with the most genetic
diversity within populations. Furthermore, the results indicated that life
form, mating pattern, gene flow, and environmental differences play a
significant role in establishing high genetic diversity of R. triflorum popu-
lations. In contrast to the results of clustering plots, analysis of environ-
mental differences demonstrated the importance of maintaining genetic
diversity among populations from different regions. Finally, the results
suggested that both in situ and ex situ conservation should be used to
protect the genetic diversity of the R. triflorum.
Sources of funding
This research was financially supported by grant K305021401, a
National Science grant from the National Science Foundation of China.
Contributions by the authors
J.J.X., B.Z., and L.Y.Z. designed the research, B.Z., and L.Y.Z. obtained
funding for the study, J.J.X., L.Y.Z., and H.F.S. collected the materials,
and J.J.X. performed the tests, analyzed the data and wrote the paper.
J.J.X. and L.Y.Z. contributed equally to the present study. All the authors
read and agreed to submit the manuscript.
Conflicts of interest statement
None declared.
Acknowledgments
The authors thank Dr. Hou-hua Li for the assistance in the laboratory
and Lin Liu for his help with the field sampling.
References
Abraha, M.T., Shimelis, H., Laing, M., Assefa, K., Amelework, B., 2016. Assessment of the
genetic relationship of tef (Eragrostis tef) genotypes using SSR markers. South
African Journal of Botany 105, 106–110.
Al-Turki, T.A., Basahi, M.A., 2015. Assessment of ISSR based molecular genetic diversity of
Hassawi rice in Saudi Arabia. Saudi Journal of Biological Sciences 22, 591–599.
Andrello, M., Henry, K., Devaux, P., Desprez, B., Manel, S., 2016. Taxonomic, spatial and
adaptive genetic variation of Beta section Beta. Theoretical and Applied Genetics
129, 257–271.
Arnaud-Haond, S., Teixeira, S., Massa, S.I., Billot, C., Saenger, P., Coupland, G., Duarte, C.M.,
Serrao, E.A., 2006. Genetic structure at range edge, low diversity and high inbreeding
in southeast Asian mangrove (Avicennia marina) populations. Molecular Ecology 15,
3515–3525.
Basahi, M., 2015. ISSR-based analysis of genetic diversity among sorghum landraces
growing in some parts of Saudi Arabia and Yemen. Comptes Rendus Biologies 338,
723–727.
Basu, T., Satya, P., Sarkar, D., Kar, C.S., Mitra, J., Karmakar, P.G., 2016. Organelle genetic
diversity in a global collection of Jute (Corchorus capsularis and C. olitorius, Malvaceae).
South African Journal of Botany 103, 54–60.
Brandao, M.M., Vieira, F.D., Nazareno, A.G., de Carvalho, D., 2015. Genetic diversity of
neotropical tree Myrcia splendens (Myrtaceae) in a fragment-corridor system in the
Atlantic rainforest. Flora 216, 35–41.
Chappell, M., Robacker, C., Jenkins, T.M., 2008. Genetic diversity of seven deciduous azalea
species (Rhododendron spp. section Pentanthera) native to the eastern United States.
Journal of the American Society for Horticultural Science 3, 374–382.
Chen, T., Huang, X.J., Zhang, J., Chen, Q., Liu, Y., Tang, H.R., Pan, D.M., Wang, X.R., 2016. Genetic
diversity and population structure patterns in Chinese cherry (Prunus pseudocerasus
Lindl) landraces. Plant Molecular Biology Reporter 34, 440–453.
Cornille, A., Gladieux, P., Smulders, M.J., Roldán-Ruiz, I., Laurens, F., Le Cam, B., Nersesyan, A.,
Clavel, J., Olonova, M., Feugey, L., 2012. New insight into the history of domesticated
apple, secondary contribution of the European wild apple to the genome of cultivated
varieties. PLoS Genetics 8, e1002703.
Cruse-Sanders, J.M., Hamrick, J.L., Ahumada, J.A., 2005. Consequences of harvesting for
genetic diversity in American ginseng (Panax quinquefolius L.), a simulation study.
Biodiversity and Conservation 14, 493–504.
Das, A., Kesari, V., Satyanarayana, V.M., Parida, A., Mitra, S., Rangan, L., 2015. Genetic diversity
in ecotypes of the scarce wild medicinal crop Zingiber moran revealed by ISSR and AFLP
marker analysis and chromosome number assessment. Plant Biosystems 149, 111–120.
Devey, D.S., Bateman, R.M., Fay, M.F., Hawkins, J.A., 2009. Genetic structure and systematic
relationships within the Ophrys fuciflora aggregate (Orchidaceae, Orchidinae), high
diversity in Kent and a wind-induced discontinuity bisecting the Adriatic. Annals of
Botany London 104, 483–495.
Dulya, O.V., Mikryukov, V.S., 2016. Genetic variation and selfing rate in Lychnisflos-cuculi
along an industrial pollution gradient. The New Phytologist 209, 1083–1095.
Dvorakova, Z., Cepkova, P.H., Janovska, D., Viehmannova, I., Svobodova, E., Cusimamani,
E.F., Milella, L., 2015. Comparative analysis of genetic diversity of 8 millet genera re-
vealed by ISSR markers. Emirates Journal of Food and Agriculture 27, 617–628.
Ellstrand, N.C., Elam, D.R., 1993. Population genetic consequences of small population
size, implications for plant conservation. Annual Review of Ecology and Systematics
24, 217–242.
Escaravage, N., Questiau, S., Pornon, A., Doche, B., Taberlet, P., 1998. Clonal diversity in a
Rhododendron ferrugineum L. (Ericaceae) population inferred from AFLP markers.
Molecular Ecology 7, 975–982.
Frankham, R., 1996. Relationship of genetic variation to population size in wildlife.
Conservation Biology 10, 1500–1508.
Geng, Y.Y., 2010. Rhododendrons notes in Linzhi. Chinese Hortic Abstracts 11 (36–38 + 197–
198 + 2).
Godt, M.J.W., Johnson, B.R., Hamrick, J., 1996. Genetic diversity and population size in four
rare southern Appalachian plant species. Conservation Biology 10, 796–805.
Hamrick, J.L., Godt, M.J.W., 1996. Conservation Genetics of Endemic Plant Species. In: Avise,
J.C., Hamrick, J.L. (Eds.), Conserv Genet. Chapman and Hall, New York, pp. 281–304.
Jena, S.N., Verma, S., Nair, K.N., Srivastava, A.K., Misra, S., Rana, T.S., 2015. Genetic diversity
and population structure of the mangrove lime (Merope angulata) in India revealed
by AFLP and ISSR markers. Aquatic Botany 120, 260–267.
Kimura, M., Crow, J.F., 1964. The number of alleles that can be maintained in a finite pop-
ulation. Genetics 49, 725–738.
Kumar, S., Ambreen, H., Murali, T.V., Bali, S., Agarwal, M., Kumar, A., Goel, S., Jagannath, A.,
2015. Assessment of genetic diversity and population structure in a global reference
collection of 531 accessions of Carthamus tinctorius L. (Safflower) using AFLP markers.
Plant Molecular Biology Reporter 33, 1299–1313.
Lande, R., Landweberg, L., Dobson, A., 1999. Extinction Risks From Anthropogenic, Ecological,
and Genetic Factors. Genetics and the Extinction of Species, DNA and the Conservation
of Biodiversity, pp. 1–22.
Larsen, B., Orabi, J., Pedersen, C., Orgaard, M., 2015. Large intraspecific genetic variation
within the saffron-crocus group (Crocus L., series crocus; Iridaceae). Plant Systematics
and Evolution 301, 425–437.
Lewontin, R., 1972. The apportionment of human diversity. Evolutionary Biology 6,
391–398.
Li, M.W., Chen, S.F., Shi, S., Zhang, Z., Liao, W.B., Wu, W., Zhou, R.C., Fan, Q., 2015. High
genetic diversity and weak population structure of Rhododendron jinggangshanicum,
a threatened endemic species in Mount Jinggangshan of China. Biochemical Systematics
and Ecology 58, 178–186.
Lienert, J., 2004. Habitat fragmentation effects on fitness of plant populations—a review.
Journal for Nature Conservation 12, 53–72.
Liu, Y.F., Xing, M., Zhao, W., Fan, R.J., Luo, S., Chen, X., 2012. Genetic diversity analysis of
Rhododendron aureum Georgi (Ericaceae) located on Changbai Mountain using ISSR
and RAPD markers. Plant Systematics and Evolution 298, 921–930.
McDermott, J.M., McDonald, B.A., 1993. Gene flow in plant patho systems. Annual Review
of Phytopathology 31, 353–373.
Mokhtari, N., Rahimmalek, M., Talebi, M., Khorrami, M., 2013. Assessment of genetic
diversity among and within Carthamus species using sequence-related amplified
polymorphism (SRAP) markers. Plant Systematics and Evolution 299, 1285–1294.
Moreira, P.D., Brandao, M.M., de Araujo, N.H., de Oliveira, D.A., Fernandes, G.W.,
2015. Genetic Diversity and Structure of the Tree Enterolobium contortisiliquum
(Fabaceae) Associated With Remnants of a Seasonally Dry Tropical Forest 210
pp. 40–46.
Morshedloo, M.R., Moghadam, M.R.F., Ebadi, A., Yazdani, D., 2015. Genetic relationships of
Iranian Hypericum perforatum L. wild populations as evaluated by ISSR markers. Plant
Systematics and Evolution 301, 657–665.
Nei, M., 1973. Analysis of gene diversity in subdivided populations. Proceedings of
the National Academy of Sciences of the United States of America 70, 3321–3323.
Ng, S.C., Corlett, R.T., 2000. Genetic variation and structure in six Rhododendron species
(Ericaceae) with contrasting local distribution patterns in Hong Kong, China. Molecular
Ecology 9, 959–969.
Ngailo, S., Shimelis, H., Sibiya, J., Amelework, B., Mtunda, K., 2016. Genetic diversity
assessment of Tanzanian sweetpotato genotypes using simple sequence repeat
markers. South African Journal of Botany 102, 40–45.
Nybom, H., 2004. Comparison of different nuclear DNA markers for estimating intraspecific
genetic diversity in plants. Molecular Ecology 5, 1143–1155.
Okcu, M., Kalkisim, O., Okcu, Z., Karabulut, B., Yildirim, N., Agar, G., 2015. Determination of
genetic diversity among wild grown apples from eastern Black Sea region in Turkey
using ISSR and RAPDs markers. Erwerbs-obstbau 57, 171–177.
Peakall, R., Smouse, P.E., 2012. GenAlEx 6.5, genetic analysis in excel. Population
genetic software for teaching and research—an update. Bioinformatics 28,
2537–2539.
 J.J. Xu et al. / South African Journal of Botany 108 (2017) 175–183
Powell, W., Morgante, M., Andre, C., Hanafey, M., Vogel, J., Tingey, S., Rafalski, A., 1996. The
comparison of RFLP, RAPD, AFLP and SSR (microsatellite) markers for germplasm
analysis. Molecular Breeding 2, 225–238.
Rafii, Z.A., Dodd, R.S., 1998. Genetic diversity among coastal and Andean natural popula-
tions of Araucaria araucana (Molina) K.Koch. Biochemical Systematics and Ecology
26, 441–451.
Rohlf, F.J., 2000. NTSYS-Pc. Numerical Taxonomy and Multivariate Analysis System,
Version 2.1. Exeter Software, Setauket, New York.
Slatkin, M., 1987. Gene flow and the geographic structure of natural populations. Science
236, 787–792.
Stout, J.C., Duffy, K.J., Egan, P.A., Harbourne, T.R., Hodkinson, T.R., 2015. Genetic diversity
and floral width variation in introduced and native populations of a long-lived
woody perennial. Aob Plants 7, plu087.
Thiel-Egenter, C., Gugerli, F., Alvarez, N., Brodbeck, S., Cieslak, E., Colli, L., Englisch, T.,
Gaudeul, M., Gielly, L., Korbecka, G., Negrini, R., Paun, O., Pellecchia, M., Rioux, D.,
Ronikier, M., Schonswetter, P., Schupfer, F., Taberlet, P., Tribsch, A., Van Loo, M.,
Winkler, M., Holderegger, R., 2009. Effects of species traits on the genetic diversity
of high-mountain plants, a multi-species study across the Alps and the Carpathians.
Global Ecology and Biogeography 18, 78–87.
Tikhonova, N.A., Polezhaeva, M.A., Pimenova, E.A., 2012. AFLP analysis of the genetic di-
versity of closely related Rhododendron species of the section Rhodorastra (Ericaceae)
from Siberia and the Far East of Russia. Russian Journal of Genetics 48, 985–992.
Trapnell, D.W., Hamrick, J.L., 2005. Mating patterns and gene flow in the neotropical epi-
phytic orchid, Laeliarubescens. Molecular Ecology 14, 75–84.
Vos, P., Hogers, R., Bleeker, M., Reijans, M., Van, D.L.T., Hornes, M., Frijters, A., Pot, J., Pelemen,
J., Kuiper, M., Zabeau, M., 1995. AFLP, a new technique for DNA fingerprinting. Nucleic
Acids Research 21, 4407–4414.
183
Wolfe, A.D., Liston, A., 1998. Contributions of PCR-based Methods to Plant Systematics
and Evolutionary Biology. In: Soltis, D.E., Soltis, P.S., Doyle, J.J. (Eds.), Plant Molecular
Systematics II. Kluwer Academic Publishers, Dordrecht, Netherlands.
Wright, S., 1951. The genetical structure of populations. Phlebologie-Annales Vasculaires
15, 323–354.
Wu, F.Q., Shen, S.K., Zhang, X.J., Wang, Y.H., Sun, W.B., 2015. Genetic diversity and population
structure of an extremely endangered species, the world's largest Rhododendron. Aob
Plants 7, plu082.
Xiao, D.M., Xuan, Y.H., Zhu, G.J., Xiong, L., 2010. Study on rapid propagation of Rhododendron
triflora. Northern Horticulture 23, 140–142.
Xing, Z., Zhang, Q.X., Liu, H., Pan, H.T., 2011. Studies on the priority exploitation ranking of
the wild ornamental plants in Mt. Shegyla. Forest Research 24, 684–693.
Xu, J.J., Zhao, B., Shen, H.F., Huang, W.M., Yuan, L.X., 2016. Assessment of genetic relation-
ship among Rhododendron cultivars using amplified fragment length polymorphism
and inter simple sequence repeat markers. Genetics and Molecular Research 15,
621–628.
Yeh, F.C., Yang, R.C., Boyle, T., Ye, Z.H., Mao, J.X., 1997. POPGENE, the User Friendly
Shareware for Population Genetic Analysis. Molecular Biology and Biotechnology
Centre. University of Alberta, Edmonton, Canada.
Yu, H.B., Zhang, L.L., 2013. Advances in phylogeography of alpine plants in the Tibetan pla-
teau and adjacent regions. Acta Botanica Boreali-Occidentalia Sinica 33, 1268–1278.
Zhang, Z.F., van Parijs, F.R.D., Xiao, B.Z., 2014. The status of AFLP in the genomics era and a
pipeline for converting AFLPs into single-locus markers. Molecular Breeding 34,
1245–1260.
Zhao, B., Yin, Z.F., Xu, M., Wang, Q.C., 2012. AFLP analysis of genetic variation in wild
populations of five Rhododendron species in Qinling Mountain in China. Biochemical
Systematics and Ecology 45, 198–205.