Genetic diversity of the world’s largest

oil palm (Elaeis guineensis Jacq.) field

genebank accessions using microsatellite
markers

C. Bakoumé, R. Wickneswari, S. Siju,

N. Rajanaidu, A. Kushairi & N. Billotte

Genetic Resources and Crop

Evolution
An International Journal

ISSN 0925-9864

Genet Resour Crop Evol

DOI 10.1007/s10722-014-0156-8

1 23
Your article is protected by copyright and all
rights are held exclusively by Springer Science
+Business Media Dordrecht. This e-offprint
is for personal use only and shall not be self-
archived in electronic repositories. If you wish
to self-archive your article, please use the
accepted manuscript version for posting on
your own website. You may further deposit
the accepted manuscript version in any
repository, provided it is only made publicly
available 12 months after official publication
or later and provided acknowledgement is
given to the original source of publication
and a link is inserted to the published article
on Springer's website. The link must be
accompanied by the following text: "The final
publication is available at link.springer.com”.

1 23
 Author's personal copy
Genet Resour Crop Evol
DOI 10.1007/s10722-014-0156-8

RESEARCH ARTICLE

Genetic diversity of the world’s largest oil palm (Elaeis

guineensis Jacq.) field genebank accessions using
microsatellite markers
C. Bakoumé • R. Wickneswari • S. Siju •
N. Rajanaidu • A. Kushairi • N. Billotte

Received: 17 March 2014 / Accepted: 22 July 2014

Ó Springer Science+Business Media Dordrecht 2014

Abstract The extent of genetic diversity among 494
oil palms from 49 populations (representing ten
African countries, three breeding materials, and one
semi-wild material) were assessed using 16 SSR
markers. The genetic diversity was high with a total of
209 alleles detected accounting for an average of 13.1
alleles per locus and a mean expected heterozygosity
of 0.644. The average genetic distance among acces-
sions was 0.769, varying from 0.000 to 1.000. Both
principal coordinates analysis and neighbor joining
tree, confirmed by structure analysis, clustered the
entire collections into three groups: the Extreme West
C. Bakoumé  R. Wickneswari (&)  S. Siju
School of Environmental and Natural Resource Sciences,
Faculty of Science and Technology, Universiti
Kebangsaan Malaysia, 43600 Bangi, Selangor, Malaysia
e-mail: wicki@ukm.edu.my; wicki@ukm.my
Present Address:
C. Bakoumé
Sime Darby Research Sdn. Bhd., KM10, Jalan Banting,
Kelenang, P.O. Box 207, 42700 Banting, Selangor,
Malaysia
Africa (EWA) group (collections from Senegal,
Guinea, and Sierra Leone), the West, Central, and
East Africa (WCEA) group (collections from Ghana,
Nigeria, Cameroon, Zaire, Angola, Tanzania, Bahia,
the semi-wild material and the two Deli breeding
materials) and the Madagascar group (collections
uniquely from Madagascar). Madagascar populations
were found to be genetically distinct from all other
African populations. The influence of human and
environmental factors might have played a major role
in grouping the African natural oil palm into three
different groups as well as in the formation of a
transition zone (formed by Ghana and Côte d’Ivoire).
Further analysis of genetic structure revealed Deli
materials as a distinct population within the WCEA
group. Given the fact that accessions were exchanged
between the EWA and WCEA groups, intra- and inter-
group combinations for breeding should be based
mainly on the genetic distance between accessions to
increase yield and heterosis.
Keywords Elaeis guineensis  Genetic diversity 
Microsatellite markers  Oil palm

N. Rajanaidu  A. Kushairi
Malaysian Palm Oil Board (MPOB),
P.O. Box 10620, Kuala Lumpur, Malaysia
Introduction
N. Billotte
UMR 1096, Centre de Coopération Internationale en
The oil palm (Elaeis guineensis Jacq. 2n = 32)
Recherche Agronomique pour le Développement
(CIRAD), TA 80/03 Avenue Agropolis, belonging to the Arecaceae family is an allogamous
34398 Montpellier Cedex 5, France aborescent monocot, cultivated extensively for the

123
 Author's personal copy
production of palm oil and palm kernel oil. It grows in
wild, semi-wild, and cultivated forms in the equatorial
tropics of Africa, Southeast Asia and South America
(Hartley 1988). Based on the presence or absence of
endocarp or shell of the fruits, three groups of E.
guineensis have been identified, viz. dura, pisifera and
tenera. The ‘Dura’ type produces thick-shelled fruit
with little pulp and is generally used as female parents
for breeding. The ‘Pisifera’ type rarely bears fruits and
does not generally contain kernels or only kernels
without shells (Mezui et al. 2013). The ‘Tenera’ type,
a hybrid of ‘Dura’ ($) and ‘Pisifera’ (#) produces
fruits with thick pulp and thin shells. Thus, ‘Tenera’
type is widely grown in commercial plantations due to
its high oil yield from the fruit mesocarp tissue
(Muniran et al. 2008).
Due to the favourable conditions, oil palm is the
most important industrial and plantation crop in
Malaysia, the largest producer and exporter of palm
oil. Majority of oil palm planting materials used for
commercial purposes are derived from Deli dura,
which was primarily introduced in Indonesia and later
into Malaysia (Corley and Tinker 2003). Palm oil
breeding is aimed at developing planting materials
with high oil yield, improved oil quality, reduced palm
height and tolerance to pests and diseases. Though
several selection cycles have resulted in yield
increases, they also have restricted the genetic base
of the initial set of populations (Rajanaidu 1994),
limiting breeding and selection within the current
breeding populations (Kushairi et al. 2003). Subse-
quently, efforts to expand the genetic base of oil palm
were initiated by exploring wild populations from the
natural range of oil palm in Africa (Rajanaidu 1994;
Rajanaidu and Jalani 1994; Hayati et al. 2004). Oil
palm collections were sampled initially from Nigeria
and further expanded to Angola, Cameroon, Gambia,
Ghana, Guinea, Madagascar, Senegal, Sierra Leone,
Zaire, and Tanzania (Rajanaidu 1994, 1995).
The Malaysian Palm Oil Board (MPOB) has an
extensive collection of E. guineensis germplasm,
maintained at Kluang Johor, Malaysia. Estimation of
genetic diversity among E. guineensis accessions is
essential for the effective characterization, utilization,
and management of genetic resources for oil palm
breeding. Molecular markers are unique in assessing
genetic diversity because they are stable and detect-
able in all tissues regardless of plant growth stage,
developmental effects, or environmental effects
123
Genet Resour Crop Evol
(Mondini et al. 2009). Though a wide range of
molecular markers have been employed for charac-
terization of oil palm accessions from Africa, Amer-
ica, and Indonesia (Shah et al. 1994; Barcelos et al.
2002; Moretzsohn et al. 2002), their application in
characterizing MPOB’s oil germplasm collections is
limited to RAPD (Rajanaidu et al. 2000a, b), AFLP
(Kularatne 2000), isozyme (Hayati et al. 2004), RFLP
markers (Rajanaidu et al. 2000a, b; Maizura et al.
2006), and SSR markers (Bakoumé et al. 2007; Singh
et al. 2008; Ting et al. 2010; Zulkifli et al. 2012).
Microsatellites or simple sequence repeats (SSRs)
are one of the most widely used molecular markers for
analysing the genetic diversity in plants due to their
multi-allelic nature, co-dominant inheritance, high
degree of polymorphism, reproducibility, locus-spe-
cific nature, and inter-lab transferability (Zane et al.
2002). A large number of samples and SSR loci can be
easily amplified and assayed for polymorphisms,
especially through multiplex PCR and/or by the mul-
tiloading of PCR products on a single gel, thus greatly
reducing the work and cost involved in the large-scale
analysis of samples (Saghai Maroof et al. 1994).
In this study, a set of 16 SSR markers (covering all
linkage groups) were used for estimating the genetic
diversity of the world’s largest oil palm field genebank
(MPOB) accessions. Though a few SSR marker based
genetic diversity analysis of MPOB germplasm col-
lections (Singh et al. 2008; Ting et al. 2010; Zulkifli
et al. 2012) have been conducted earlier, the present
study has its own particular significance as it generated
a much more accurate picture on genetic structure of
the wild palms according to regions rather than
population or countries using more number of acces-
sions and informative markers. The estimated genetic
distance will also help in understanding the genetic
relatedness of oil palm accessions for adopting
appropriate sampling strategies for future collection
and conservation of exotic germplasm, beyond being
exploited for crop improvement purposes.
Materials and methods
Plant materials and DNA extraction
A total of 464 leaf samples, representing 45 natural
populations (Table 1) maintained at the field gene-
bank of the Malaysian Palm Oil Board (MPOB) at
 Author's personal copy
Genet Resour Crop Evol

Table 1 Distribution of common and rare alleles among oil palm accessions collected from different localities of each country and
breeding materials as revealed by 16 SSR markers
Country Pop. Locality No. of families No. of palms Total alleles Common Rare
p C 0.05 p \ 0.05

Senegal 5 Senegal 1a 6 14 90 73 17
8 Takene 4 11 68 48 20
13 Senegal 2a 5 11 60 45 15
Guinea-Conakry 1 Korodu 1 10 94 94 0
4 Seredou 4 12 96 66 30
12 Kastiri 4 11 85 64 21
Sierra Leone 1 York 3 11 83 56 27
5 Bonyeya 3 11 83 63 20
13 Masenie 3 12 96 69 27
Ghana 3 Adidi 3 11 83 66 17
9 Mpohor/Ayiem 2 10 100 100 0
13 Dumkofo/Nkekasu 2 10 94 94 0
Nigeria 12 Umuabi 7 11 94 77 17
35 Ibono 4 9 81 81 0
40 Vabiti 8 9 107 107 0
44 Ogbaloto 8 12 97 71 26
45 Ologba 7 12 105 71 34
Cameroon 1 Eyam 2 10 70 70 0
2 Akwen 2 10 84 84 0
9 Widikum 11 11 93 70 0
13 Kikangko 2 11 80 64 16
16 Bafang 1 10 78 78 0
22 Bertoua 1 10 68 68 0
28 Bipindi 1 10 72 72 0
29 Ngambe 5 10 94 94 0
30 Ndian 3 10 82 82 0
31 Bafut 2 10 88 88 0
Zaire 7 Mobila 4 10 90 90 0
22 Isangi 5 10 81 81 0
29 Mbanza Ngungu 5 10 88 88 0
30 Kiringye 2 10 86 86 0
36 Luki 3 10 77 77 0
Angola 1 Bengo 5 10 91 91 0
3 Funda 7 10 88 88 0
5 Sumbe 2 10 88 88 0
6 Benguela 6 10 68 68 0
8 Santa Catarina 9 11 90 66 24
Tanzania 1 Ujiji 5 10 91 91 0
3 Kiganza 5 10 99 99 0
5 Kwitanga Government Farm 5 10 87 87 0
7 Chankabwimba 5 11 89 68 21
9 Simbo village 2 10 83 83 0

123
 Author's personal copy
Genet Resour Crop Evol

Table 1 continued
Country Pop. Locality No. of families No. of palms Total alleles Common Rare
p C 0.05 p \ 0.05

Madagascar 1 Malaimbandy 1 6 11 18 18 0
3 Malaimbandy 2 3 5 17 17 0
4 Malaimbandy 3 1 7 18 18 0
Deli MPOB 1 Breeding material 9 11 65 53 12
Deli Dabou – Breeding material NA 9 56 56 0
La Mé – Bingerville 2 6 57 57 0
Bahia – Bahia NA 4 49 49 0
a
Locality name not documented

SENEGAL

13 GUINEA NIGERIA
5 8
12
13 1 40 45
1 5 4 13 44
3 12
35
13 31
9
2 1 9 16
30 29
SIERRA LEONE 28 22
GHANA 7
22
CAMEROON ZAIRE
29 30
7
36 3 TANZANIA
8 9
1
15
3
5
6
ANGOLA

4
3
1

0 350 700 miles

0 350 700 kilometers
MADAGASCAR

(......): limit of the distribution area; (•): population sampled in this study

Fig. 1 Map of Africa showing the location of oil palm material sampling sites. The assigned population number and locality of
collection within each country is mentioned in Table 1

Kluang Johor, Malaysia were used in the study. (Côte d’Ivoire), and La Mé (Côte d’Ivoire) breeding
The analyzed samples include accessions originat- materials and a semi-wild material Bahia (Brazil).
ing from 10 African countries (Fig. 1) along with Genomic DNA was extracted from fresh leaf
Deli dura MPOB (Malaysia), Deli dura Dabou tissues (leaflets of an unopened leaf from the spear)

123
 Author's personal copy
Genet Resour Crop Evol
using CTAB method (Doyle and Doyle 1990) and the
DNA concentration was determined using agarose gel
electrophoresis.
Microsatellite analysis
A set of 16 E. guineensis highly polymorphic SSR loci
representing each linkage groups (mEgCIR0173, mEg-
CIR0353, mEgCIR0802, mEgCIR0832, mEgCIR1730,
mEgCIR1753, mEgCIR3282, mEgCIR3292, mEg-
CIR3300, mEgCIR3362, mEgCIR3363, mEgCIR3543,
mEgCIR3546, mEgCIR3574, mEgCIR3785, and mEg-
CIR3886) developed by Billotte et al. (2005) were used
for genotyping. PCR amplification was carried out in a
thermal cycler (Master Cycler EP Gradient S, Eppen-
dorf, Germany) as described by Billotte et al. (2001).
The amplified products were resolved in a 7 %
denaturing polyacrylamide gel (20 cm 9 16 cm) using
Protean II xi cell (Bio-Rad, Hercules, USA) and were
silver stained (Creste et al. 2001). To limit the number
of electrophoresis gels used, PCR products of two to
four SSR loci (showing a reference allele size differ-
ence of C30 bp) were mixed and resolved on the same
gel (multiple loading). Genomic DNAs from ZA21 and
ZA29, two accessions from Zaire (Democratic Repub-
lic of Congo; DRC), were used as controls for each gel
along with two lanes for 10 bp size ladder to facilitate
the allele scoring for a given gel and to link one gel to
another for the same locus.
Bands were scored by size, not by the homology of
the sequences. A counter scoring was carried out in
two different laboratories (Plant Genetics Laboratory
of Universiti Kebangsaan Malaysia and Biotrop
Laboratory of CIRAD-France), which validated the
alleles and the subsequent genotype data generated.
Data analysis
Basic statistics such as allelic frequency, mean number
of alleles per locus (A), effective number of alleles
(Ae; Crow and Kimura 1970), observed heterozygosity
(Ho), fixation indices (Fis), and genetic distances
among accessions (Nei and Takezaki 1983) were
calculated using PowerMarker version 3.25 (Liu and
Muse 2005). The corrected expected heterozygosity
(He) for small and unequal sample sizes (Nei 1978)
and their variances were calculated with POPGENE
version 1.32 (Yeh and Boyle 1999). Chi square tests
for each locality were performed using Minitab
software. The approach of Marshall and Brown
(1975) was adopted for obtaining a better comparison
of the distribution of common alleles (alleles at
frequency p C 0.05) and rare alleles (alleles at
frequency p \ 0.05).
A principal coordinates analysis (PCoA) was
performed and a neighbour-joining tree (NJTree)
was constructed based on the dissimilarity matrix
using DARwin5 software (Perrier and Jacquemoud-
Collet 2006) to visualize the main provenances of the
natural oil palm collections studied and for a better
readability of the relatedness between accessions. For
the dissimilarity matrix, dissimilarity indices between
accessions i and j (dij) were estimated with 1,000
bootstraps by simple matching as follows:
1X L
ml
dij ¼ 1 
L l¼1 p
where L is the number of loci, p the ploidy, and ml the
number of common alleles between i and j (for locus
l).
Because the 16 SSR markers were unlinked, the
programme STRUCTURE version 2.2 (Pritchard et al.
2000), with burning period and repetitions set at
10,000 and 100,000, respectively, was used for
identifying the organisation of populations in k groups
and for the allocation of accessions to different
populations.
Results
Allelic diversity and genetic variation in E.
guineensis accessions
A total of 209 alleles were detected among the 494 oil
palm genotypes. The number of alleles per SSR locus
ranged from 8 (mEgCIR1730 and mEgCIR1753) to 22
(mEgCIR3362), with an average of 13.1 alleles per
SSR loci. Rare alleles (frequency p \ 0.05) were
found in almost all the localities from Senegal,
Conakry Guinea, and Sierra Leone, which are neigh-
bouring countries with dry weather. Higher numbers
of rare alleles were found in the populations of Ologba
(population 45) and Seredou (population 4) (Table 1).
The estimates for genetic diversity parameters for the
complete set of oil palm collections are given in
Table 2.
123
 Author's personal copy
Genet Resour Crop Evol

Table 2 Genetic diversity parameter estimates and fixation Table 2 continued

index for 49 natural populations and breeding materials of E.
guineensis, based on 16 microsatellite loci Country Pop. N A Ae Ho He Fis

Country Pop. N A Ae Ho He Fis Madagascar 1 11 1.2 1.1 0.0340 0.041 0.190*

3 5 1.1 1.1 0.025 0.033 0.138*
Senegal 5 14 5.2 3.0 0.542 0.685 0.124
4 7 1.1 1.1 0.000 0.055 1.000*
8 11 4.3 2.4 0.398 0.556 0.219*
Deli Dabou 8 3.5 2.2 0.310 0.549 0.336*
13 11 3.8 2.2 0.352 0.526 0.254*
Deli MPOB 1 11 4.1 2.3 0.211 0.559 0.569*
Guinea- 1 10 5.8 3.0 0.592 0.701 0.049
La Mé 6 3.6 2.2 0.531 0.618 -0.054
Conakry 4 12 5.9 3.2 0.485 0.722 0.251*
Bahia 4 3.3 2.0 0.396 0.668 0.116
12 11 5.4 3.0 0.449 0.688 0.271*
Mean 5.0 2.8 0.460 0.644 0.206
Sierra Leone 1 11 5.2 2.8 0.450 0.688 0.252*
5 11 5.2 2.9 0.542 0.672 0.09 N Number of samples per population, A mean number of alleles
per locus, Ae effective number of alleles per locus, Ho observed
13 12 6.0 3.2 0.480 0.665 0.274* heterozygosity, He expected heterozygosity, Fis fixation index
Ghana 3 11 5.3 2.9 0.523 0.686 0.153* * Significantly different from zero (p B 0.05)
9 10 6.2 3.2 0.575 0.741 0.129
13 10 5.9 3.2 0.550 0.729 0.153*
Based on a country-wise description, the average
Nigeria 12 11 5.9 3.2 0.529 0.736 0.199*
number of alleles per locus ranged from 1.1 (Malaim-
35 9 5.1 2.9 0.516 0.712 0.169*
bandy 1 and 2, Madagascar) to 6.7 (Vabiti, Nigeria)
40 9 6.7 3.6 0.543 0.803 0.224*
with a mean value of 4.98. The effective number of
44 12 6.1 3.3 0.551 0.751 0.191* alleles ranged from 1.1 (all Madagascar localities)
45 12 6.6 3.5 0.526 0.739 0.215* to 3.6 (Vabiti, Nigeria and Kiganza, Tanzania)
Cameroon 1 10 4.4 2.4 0.375 0.637 0.333* (mean = 2.78). The observed heterozygosity (Ho)
2 10 5.3 3.1 0.551 0.689 0.105* among the 49 populations was lowest in Malaimbandy
9 11 5.9 3.2 0.465 0.739 0.297* 1, Madagascar (Ho = 0.000) and highest in Kiganza,
13 11 4.9 2.7 0.413 0.651 0.289* Tanzania (Ho = 0.615), with an average of 0.460. The
16 10 4.9 2.6 0.578 0.650 0.007 expected heterozygosity (He) ranged from 0.033
22 10 4.3 2.3 0.412 0.566 0.181 (Malaimbandy 2, Madagascar) to 0.803 (Vabiti,
28 10 4.5 2.5 0.500 0.627 0.106 Nigeria) (mean = 0.64). The Student test showed that
29 10 5.9 3.2 0.494 0.727 0.234* Fis values were mostly positive for all loci with mean
30 10 5.1 2.8 0.463 0.678 0.230* Fis values ranging from 0.007 (Kafang, Cameroon) to
31 10 5.5 2.8 0.432 0.644 0.242* 1.00 (Malaimbandy 3, Madagascar), with an average
Zaire 7 10 5.6 3.0 0.469 0.704 0.298* value of 0.207. The only negative Fis value (-0.054)
22 10 5.1 3.0 0.535 0.698 0.138 was found in La Mé materials.
29 10 5.5 2.9 0.470 0.688 0.228*
30 10 5.3 3.0 0.483 0.724 0.242* Genetic relatedness among E. guineensis
36 10 4.8 2.6 0.461 0.668 0.209* accessions
Angola 1 10 5.6 3.0 0.508 0.690 0.171*
3 10 5.5 2.8 0.494 0.649 0.146* The average genetic distance among accessions was
5 10 5.5 3.1 0.588 0.700 0.062 0.769, varying from 0.000 to 1.000. The value of
6 10 4.3 2.5 0.544 0.636 0.044 genetic distance among Madagascar accessions was
8 11 5.6 3.0 0.460 0.714 0.278* the lowest and ranged from 0.000 to 0.125. The
Tanzania 1 10 5.7 3.2 0.503 0.734 0.228* genetic distance values showed no trend among
3 10 6.2 3.6 0.615 0.770 0.105 accessions from the same locality or the same country.
5 10 5.4 3.2 0.594 0.730 0.089 No complete divergence (D = 1.000) was observed
7 11 5.6 3.3 0.489 0.745 0.269* among accessions of the same country except for two
9 10 5.2 3.0 0.540 0.700 0.133* cases: between accession NI43 from Ayangba and
accession NI33 from Bida in Nigeria, and between

123
 Author's personal copy
Genet Resour Crop Evol
Fig. 2 Principal coordinate
analysis based on
dissimilarity matrix using 16
SSR markers
West, Central and East Africa
and breeding materials
accession ZA02 from Mobila and accession ZA38
from Kiringye in the Democratic Republic of Congo
(data not shown).
The PCoA clustered the localities or locality-derived
collections into three groups (Fig. 2). All collections
from Madagascar formed the Madagascar (M) group,
while collections from Senegal, Guinea, and Sierra
Leone formed the Extreme West Africa (EWA) group.
Collections from Ghana, Nigeria, Cameroon, Zaire,
Angola, Tanzania, Deli MPOB, Deli Dabou, La Mé
breeding materials, and Bahia semi-wild material con-
stituted the West-Central-East Africa (WCEA) group. A
clear separation of Madagascar accessions from the
remaining group was evident from this analysis.
The NJTree separated the WCEA group from the
EWA group (Fig. 3). It revealed a better picture of the
group’s structure, which was clouded to a certain
extent by the admixture of accessions from different
localities and countries in the data subjected to PCoA
analysis. The Madagascar accessions maintained a
separate grouping pattern, a pattern similar to that of
PCoA analysis. The Deli breeding materials clustered
together and were well defined in the WCEA group.
Accessions from Cameroon and Nigeria were mostly
scattered within the WCEA group. A very small
number of accessions from La Mé and Bahia prove-
nances were also mixed with other accessions in the
WCEA group. Accessions of Guinea, Sierra Leone,
and Senegal provenances were clearly demarcated in
the EWA group.
Madagascar
Extreme West Africa
Genetic structure of oil palm collections
The organization of oil palm collections into three
different groups of populations was the most compre-
hensive (Fig. 4). The grouping was similar to those
obtained in the PCoA and NJTree analysis. The
collections from Guinea, Sierra Leone and Senegal
formed one group (EWA), while collections from
Ghana, Nigeria, Cameroon, Democratic Republic of
Congo, Angola and Tanzania constituted the WCEA
group. Breeding and semi-wild materials clustered once
more with the collections from WCEA group. All
collections from Madagascar maintained a distinct
grouping pattern. The extent of genetic relatedness
between accessions based on origin was revealed by
altering the k value. When k was increased from k = 3
to k = 4 and k = 5, the Madagascar group remained
intact while some accessions from the EWA group were
allocated to the WCEA group and vice versa. About
10 % of the accessions from the WCEA group were
allocated to the EWA group. At k = 5, Deli Dabou and
Deli MPOB materials formed a distinct group.
Discussion
The present study detected a total of 209 alleles across
16 SSR loci, with an average of 13.1 alleles/locus.
This is slightly less than the average number of alleles
(14.5 alleles/locus) revealed by SSRs used for
123
 Author's personal copy
Guinea
Extreme West Africa Senegal
Sierra Leone
Deli materials
Ghana
Nigeria
Zaire
Fig. 3 Neighbor-joining tree drawn from the dissimilarity matrix generated by16 SSR markers genotyped across the oil palm
collections (494 oil palms from 49 populations). A scale for genetic distance is provided at the bottom of the graph
characterizing germplasm collections from Asia,
Africa, and Latin America (Cochard et al. 2009).
The variation in number of alleles detected might be
due to the variation in the number of accessions
sampled and origin of the samples from each popu-
lation. However, the application of other molecular
markers such as isozymes (Hayati et al. 2004) and
RFLP (Maizura et al. 2006) revealed a much lower
average number of alleles (1.8 alleles/locus) in the
same set of oil palm collections. These findings
indicate the robustness of SSR markers in revealing
higher numbers of alleles than isozymes/allozymes, a
fact that was also reported earlier (Bruford et al. 1996).
Allelic frequencies at each locus varied from
locality to locality with certain alleles confined only
to one or limited localities within a country. The
variation in allele frequencies among the populations
might be probably due to genetic drift, natural
selection and human intervention (Bakoume et al.
2007). Mutation alone creates new alleles and is a
suggestive force in increasing genetic variation (Klug
123
Genet Resour Crop Evol
Madagascar
Tanzania
Angola
West, Central and East Africa
Cameroon
and breeding materials
and Cummings 1991). Among the samples analyzed,
accessions from all the Madagascar localities reported
the lowest average number of alleles (1.1). Rare alleles
were completely absent in Madagascar populations.
The oil palm groves on the island are isolated and
considered to be small (Rajanaidu et al. 2000a). A high
degree of inbreeding among small populations over a
sufficient number of generations could have resulted
in fixing some alleles.
The prevalence of more rare alleles (p \ 0.05) in
populations from Senegal, Guinea, and Sierra Leone
(neighbouring countries with low rainfall and dry
weather) may be related to adaptive traits, resulting
from mutation. These rare alleles would not have
spread across the entire African oil palm belt due to
their isolation by distance from natural groves in other
countries. However, the presence of the same rare
alleles across localities from different countries with
similar dry weather conditions also suggests that their
presence might be not only due to mutation but also
due to selection. Similar instances were revealed using
 Author's personal copy
Genet Resour Crop Evol
AN CM GH GU MA
AN CM GH GU MA
AN CM GH GU MA
Fig. 4 Population structure analysis of 494 oil palm accessions
based on microsatellite marker analysis. K-values of 3–5 sub-
populations are shown to right and naming of oil palm
populations are as follow—AN Angola, CM Cameroon, GH
isozymes (Hayati et al. 2004), which was attributed
due to genetic drift and adaptive genetic variants.
He value was highest for population number 40
from Nigeria (He = 0.803), and lowest in population
number 3 from Madagascar (He = 0.033). Reproduc-
tive isolation favoring inbreeding among limited
accessions might be the possible reason for the
reduced expected heterozygosity level in Madagascar
populations (Zulkifli et al. 2012). The comparatively
high genetic diversity of populations from Nigeria
(evidenced by high values for He and average number
of alleles) also indicates it to be the possible centre of
diversity of oil palm (Hayati et al. 2004; Maizura et al.
2006; Zulkifli et al. 2012). The mean He value for the
analyzed samples was high (He = 0.644) and is
comparable with the values reported by Billotte et al.
(2001) (He = 0.68). However, Cochard et al. (2009)
reported a slightly increased mean He value of 0.704.
The high genetic diversity observed might have
resulted from the out-crossing behaviour of oil palm,
which might explain its plasticity in adaptation to
various environments and large distribution area,
particularly within the dry and cold tropics of Africa.
The Fis values were mostly positive and considerably
different from zero (p \ 0.05), revealing a deficiency
k =3
NI SN SL TA ZA DM LM DDBA
k =4
NI SN SL TA ZA DMLMDD BA
k =5
NI SN SL TA ZA DMLMDDBA
Ghana, GU Guinea-Conakry, MA Madagascar, NI Nigeria, SN
Senegal, SL Sierra Leone, TA Tanzania, ZA Zaire, DM Deli
MPOB, LM La Mé, DD Deli Dabou, BA Bahia
of heterozygotes, which suggests deviation from the
Hardy–Weinberg equilibrium (HWE), mostly due to
inbreeding.
The average genetic distance among accessions
was rather high. In oil palm, almost 70 % of the
genetic variation exists among individuals within
populations (Kularatne 2000; Hayati et al. 2004;
Cochard et al. 2009). The oil palm inflorescences are
composed of thousands of flowers, pollinated mostly
by insects and to a small extent by wind. Hence, it is
believed that many palms are pollen donors for a given
inflorescence, leading to genotypic differences among
oil palms derived from the same bunch.
The PCoA and NJTree resolved the entire acces-
sions to three distinct groups—Madagascar (M) group,
West-Central-East Africa (WCEA), and Extreme
West Africa (EWA) group, the composition of which
was also confirmed by genetic structure analysis. The
grouping pattern is in agreement with the geographical
locations and is similar to those identified using AFLP
(Kularatne 2000), isozymes (Hayati et al. 2004), and
RFLP markers (Maizura et al. 2006).
A systematic separation of Madagascar (M) group
from other African collections is evident in all
analysis. As Madagascar is separated from mainland
123
 Author's personal copy
Africa by the Mozambique Channel, restricted pollen
movement and seed dispersal across the geographic
barrier might have limited the gene flow from the
mainland, resulting in an independent evolution of a
distinct population with unique features. Oil palm is a
marginal oil crop in Madagascar with small bunches
with small fruits and low yield for palm oil. It is of
little or no interest to the people of the island and
consequently is subjected to little human disturbance.
Hence, oil palm has experienced limited expansion in
the island and is confined to the western part of
Madagascar between Cape Saint André and the
Fiheranana as far south as latitude 21°S (Zeven
1967) with limited populations (Rajanaidu et al.
2000a).
The WCEA group, on the other hand, constitute
countries where oil palm is regarded as a staple source
of lipids for human consumption. Human interven-
tions might have resulted in the movement of genetic
material across different localities, resulting in a rapid
expansion of genetic diversity. Human interventions
within the group include (1) selection and preservation
of only oil palms with good, fresh fruit yields, (2)
felling the low or unproductive oil palms and/or use of
their inflorescences for the production of palm wine,
and (iii) collection of palm from remote localities for
the production of palm oil and subsequent trading to
different places.
The clustering of the Brazilian Bahia semi-wild
materials in the WCEA group confirms their geo-
graphical origin. In fact, Brazilian groves were
established during the period of the slave trade on a
belt 5–8 km wide from the Bay of Salvador to Maraú
(Bahia state). Hartley (1988) reported that the Brazil-
ian groves were very similar in composition and yield
of fruits with those from southeastern Nigeria. How-
ever, Cochard et al. (2009) revealed that the La Mé
(Côte d’Ivoire) materials form a different group from
that of Ghana, Nigeria, Cameroon, Zaire, Angola, and
Tanzania.
The allocation of accessions to the three popula-
tions showed that Côte d’Ivoire (represented by the La
Mé materials) and Ghana could constitute a transition
zone between the EWA and WCEA, given that
considerable proportion of accessions were exchanged
with Guinea Conakry, Sierra Leone, and Senegal. The
inclusion of Deli materials within the WCEA group is
also suggestive of their possible origin from this
African region. Studies using DNA-based markers
123
Genet Resour Crop Evol
(Rajanaidu et al. 2000a, b; Kularatne 2000) and
protein markers (Ghesquiere 1985; Hayati et al. 2004)
have also suggested the possible origin of Deli
materials in a region covering Nigeria, Cameroon,
Zaire, and Angola.
In EWA countries, oils from groundnut and cotton
are the staple sources of lipids, whereas palm oil is
marginal in local diets. The climate of Guinea
Conakry, Sierra Leone, and Senegal are characterized
by dry weather and long dry season. These climatic
features could have induced adaptation in oil palms,
resulting in the creation of alleles unique to these
countries. The introduction of oil palm into the EWA
region could have been relatively recent when com-
pared with that of countries of the WCEA group. In
fact, mention of oil palm groves in Liberia was made
in the early sixteenth century (Hartley 1988; Corley
and Tinker 2003).
When the k was increased to 5, more exchanges of
materials among countries were revealed (Fig. 4).
Furthermore, Deli materials formed a different group.
Although Deli materials were derived from four
seedlings introduced from Africa to the Bogor Botanic
Garden in Java, Indonesia in 1848 (Hartley 1988),
selection carried out over more than one century
(starting with mass selection) might have resulted in
the formation of a population distinct from those of the
African dura. Hence, the WCEA group comprised
sub-populations, i.e., breeding materials including La
Mé materials and Bahia semi-wild materials. The
differences in clustering (k = 5) warns us to be
flexible when applying the notion of grouping of oil
palm accessions on the basis of country for selection or
conservation purposes.
In oil palm, good complementarity of yield com-
ponents and heterosis are the main factors involved in
ensuring a high yield. The heterosis effect represents
15–30 % of the mid-parental value for total bunch
production, 3–5 % for oil to mesocarp ratio, and
16–20 % for oil extraction rate (Noiret et al. 1966;
Gascon et al. 1969) whilst inbreeding reduces bunch
yield as much as 50 % (Gascon et al. 1969) and
depresses vegetative characters and bunch production
components (Cao 1995).
It is clearly evident from SSR analysis that genetic
distance among the accessions plays a major role in
preventing inbreeding depression in the intra or inter-
group progenies. Based on earlier observations, dura
and tenera from Nigerian collections at MPOB were
 Author's personal copy
Genet Resour Crop Evol
introgressed into the Deli dura group and African
tenera/pisifera group, respectively (Rajanaidu 2013;
personal communication). The good performance
(high fresh fruit bunch yield) of the resulting progenies
correlated with the high genetic distance among the
parents, as revealed in this study. The same factor
might also be responsible for the superior performance
of MPOB’s intra-group progenies viz., Cameroon 9
Cameroon and Nigeria 9 Nigeria. These intra-group
crosses were reported to produce highest number of
bunches (18.2/palm/years) and shortest palms (1.8 m
or 47 % of the control), respectively (Amiruddin et al.
2005). By correlating the genetic distance observed
between these accessions and the comparative perfor-
mance of progenies, it is clearly apparent that the
breeding strategy followed by MPOB to carryout
intra-group crosses (rather than intergroup crosses)
using genetically distant accessions often culminates
in enhanced heterosis.
To conclude, the high genetic diversity identified
within the MPOB’s African oil palm germplasm
collections provides ample scope for further genetic
improvement. The grouping of the MPOB collections
into three distinct populations reflects and confirms
the history of species distribution and sociocultural
management of oil palm within the African oil palm
belt. Human and environmental factors might have
contributed to the cataloguing of African natural oil
palm into three different populations and the forma-
tion of a transition zone (formed by Ghana and Côte
d’Ivoire). Madagascar populations were found to be
genetically distinct from other African collections in
all analyses. The study also revealed the existence of
high genetic diversity in Nigerian oil palm popula-
tions supporting it to be the possible centre of
diversity of E. guineensis. Estimation of the extent of
genetic distance among accessions should be given
paramount importance for future oil palm breeding
programmes, in order to enhance yield performances
by maximising the heterosis effect and minimising
inbreeding depression.
Acknowledgments The present study was carried out in the
Plant Genetics Laboratory of Universiti Kebangsaan Malaysia
(UKM) with the joint support of the Malaysian Palm Oil Board
(MPOB) in Malaysia, the Centre de Coopération Internationale
de Recherche Agronomique pour le Développement (CIRAD)
in France, and the Institute of Agricultural Research for
Development (IRAD) in Cameroon. Our sincere thanks to
Alena Sanusi for helpful comments on the manuscript.
References
Amiruddin MD, Kushairi A, Rajanaidu N, Noh A, Isa ZA (2005)
Performance of various oil palm introgressed populations
at MPOB. Proceedings of the international conference on
technological breakthroughs and commercialization—the
way forward. Malaysian Palm Oil Board, Petaling Jaya,
pp 111–143
Bakoumé C, Wickneswari R, Rajanaidu N, Kushairi A, Amblard
P, Billotte N (2007) Allelic diversity of natural oil palm
(Elaeis guineensis Jacq.) populations detected by micro-
satellite markers: implications for conservation. Plant
Genet Resour C&U 5:104–107
Barcelos E, Amblard P, Berthaud J, Seguin M (2002) Genetic
diversity and relationship in American and African oil
palm as revealed by RFLP and AFLP molecular markers.
Pesqui Agropecu Bras 37:1105–1114
Billotte N, Risterucci AM, Barcelos E, Noyer L, Amblard P,
Baurens FC (2001) Development, characterization, and
across-taxa utility of oil palm (Elaeis guineensis Jacq.)
microsatellite markers. Genome 44:413–425
Billotte N, Marseillac N, Risterucci AM, Adon B, Brottier P,
Baurens FC, Singh R, Herrán A, Asmady Billot C, Amb-
lard P, Durand-Gasselin T, Courtois B, Asmono D, Cheah
SC, Rohde W, Ritter E, Charrier A (2005) Microsatellite-
based high density linkage map in oil palm (Elaeis guine-
ensis Jacq.). Theor Appl Genet 110:754–765
Bruford MW, Cheesman DJ, Coote T, Green AA, Haines SA,
O’Ryan C, Williams TR (1996) Microsatellites and their
application to conservation genetics. In: Smith TB, Wayne
RK (eds) Molecular genetic approaches in conservation.
Oxford University Press, Oxford, pp 278–297
Cao TV (1995) Organisation de la variabilité génétique chez le
palmier à huile Elaeis guineensis Jacq. Conséquences pour
l’amélioration des populations et la création variétale.
Ph.D. thesis, Institut National Agronomique, Paris
Cochard B, Adon B, Rekima S, Billotte N, De Chenon RD,
Koutou A, Nouy B, Omoré A, Purba AR, Glazsmann JC,
Noyer JL (2009) Geographic and genetic structure of
African oil palm diversity suggests new approaches to
breeding. Tree Gen Genom 5:493–504
Corley RHV, Tinker PB (2003) The oil palm, 4th edn. Blackwell
Science Ltd, Oxford
Creste S, Tulmann N, Figueira A (2001) Detection of single
sequence repeat polymorphism in denaturing polyacryl-
amide sequencing gels by silver staining. Plant Mol Biol
Rep 19:200–306
Crow JF, Kimura M (1970) An introduction to population
genetic theory. Harper and Row, New York
Doyle J, Doyle L (1990) Isolation of plant DNA from fresh
tissue. Focus 12:13–15
Gascon JP, Noiret JM, Meunier J (1969) Effets de la consan-
guinité chez Elaeis guineensis Jacq. Oleagineux
24:603–607
Ghesquiere M (1985) Enzyme polymorphism in oil palm (Elaeis
guineensis Jacq.). II. Variability and genetic structure of
seven origins of oil palm. Oleagineux 40:529–540
Hartley CWS (1988) The oil palm, 2nd edn. Longman, London
Hayati A, Wickneswari R, Maizura I, Rajanaidu N (2004)
Genetic diversity of oil palm (Elaeis guineensis Jacq.)
123
 Author's personal copy
germplasm collections from Africa: implications for
improvement and conservation of genetic resources. Theor
Appl Genet 108:1274–1284
Klug WS, Cummings MR (1991) Concepts of genetics, 3rd edn.
Macmillan, Upper Saddle River, New Jersey
Kularatne RS (2000) Assessment of genetic diversity in natural
oil palm (Elaeis guineensis Jacq.) using amplified fragment
length polymorphism markers. Ph.D. Thesis, Universiti
Kebangsaan Malaysia, Bangi, Malaysia
Kushairi A, Rajanaidu N, Mohd Din A (2003) Mining the
germplasm. In: Seminar on the progress of oil palm
breeding and selection, Medan, Sumatra
Liu K, Muse SV (2005) Powermarker: integrated analysis
environment for genetic analysis. Bioinformatics
21:2128–2129
Maizura I, Rajanaidu N, Zakri AH, Cheah SC (2006) Assess-
ment of genetic diversity in oil palm (Elaeis guineensis
Jacq.) using restriction fragment length polymorphism
(RFLP). Genet Res Crop Evol 53:187–195
Marshall DR, Brown AHD (1975) Optimum sampling strategy
in genetic conservation. In: Frankel OH, Hawkes JG (eds)
Crop genetic resources for today and tomorrow. Cam-
bridge University Press, London, pp 53–80
Mezui MRA, Tchouamo IR, Baudouin M (2013) Adoption of
the tenera hybrid of oil palm (Elaeis guineensis Jacquin)
among smallholder farmers in Cameroon. Tropicultura
31:103–109
Mondini L, Noorani A, Pagnotta MA (2009) Assessing plant
genetic diversity by molecular tools. Diversity 1:19–35
Moretzsohn MC, Ferreira MA, Amaral ZPS, Coelho PJA,
Grattapaglia D (2002) Genetic diversity of Brazilian oil
palm (Elaeis oleifera H.B.K.) germplasm collected in the
Amazon Forest. Euphytica 124:35–45
Muniran F, Bhore SJ, Shah FH (2008) Micropropagation of
Elaeis guineensis Jacq. ‘‘Dura’’: comparison of three basal
media for efficient regeneration. Indian J Exp Biol
46:79–82
Nei M (1978) Estimation of average heterozygosity and genetic
distance from a small number of individuals. Genetics
89:583–590
Nei M, Takezaki N (1983) Estimation of genetic distances and
phylogenetic trees from DNA analysis. Proc 5th World
Congr Genet Appl Livest Prod 21:405–412
Noiret JM, Gascon JP, Bénard G (1966) Contribution à l’étude
de l’hérédité des caractéristiques de la qualité du régime et
du fruit d’Elaeis guineensis Jacq. Oleagineux 21:343–349
Perrier X, Jacquemoud-Collet JP (2006) DARwin software.
http://darwin.cirad.fr/darwin
Pritchard JK, Stephens M, Donnelly P (2000) Inference of
population structure using multilocus genotype data.
Genetics 155:945–959
123
Genet Resour Crop Evol
Rajanaidu N (1994) PORIM oil palm gene bank. Palm Oil
Research Institute, Bangi, Malaysia
Rajanaidu N (1995) PORIM oil palm genebank. Palm Oil Tech
Bull 1:7
Rajanaidu N, Jalani BS (1994) Oil palm genetic resources
collection, evaluation, utilization and conservation. In:
PORIM Colloquium on Oil Palm Genetic Resources.
PORIM, Bangi, Malaysia
Rajanaidu N, Maizuram I, Cheah SC (2000) Screening of oil
palm natural populations using RAPD and RFLP molecular
markers. In: Rajanaidu N, Ariffin D (eds) Proceedings of
international symposium on oil palm genetic resources and
utilization, Kuala Lumpur, pp AA1–AA28
Rajanaidu N, Kushairi A, Rafii M, Mohd Din A, Maizura I, Isa
ZA, Jalani BS (2000a) Oil palm genetic resources and their
utilization: A review. In: Rajanaidu N, Ariffin D (eds)
Proceedings of international symposium on oil palm
genetic resources and utilization, Kuala Lumpur, pp A1–
A55
Saghai Maroof MA, Biyashev RM, Yang GP, Zhang Q, Allard
RW (1994) Extraordinary polymorphic microsatellite
DNA in barley: species diversity, chromosomal locations,
and population dynamics. Proc Natl Acad Sci USA
91:5466–5470
Shah FH, Rashid O, Simons AJ, Dunsdon A (1994) The utility of
RAPD markers for the determination of genetic variation in
oil palm (Elaeis guineensis). Theor Appl Genet
89:713–718
Singh R, Zaki NM, Ting NC, Rosli R, Tan SG, Low ETL, Ithnin
M, Cheah SC (2008) Exploiting an oil palm EST database
for the development of gene-derived SSR markers and their
exploitation for assessment of genetic diversity. Biologia
63:227–235
Ting NC, Noorhariza MZ, Rozana R, Low ETL, Ithnin M,
Cheah SC, Tan SG, Singh R (2010) SSR mining in oil palm
EST database: application in oil palm germplasm diversity
studies. J Genet 89:135–145
Yeh FC, Boyle T (1999) POPGENE Version 1.32. The user-
friendly software for population genetic analysis. Univer-
sity of Alberta and CIFOR, Calgary, Alta
Zane L, Bargelloni L, Patarnello T (2002) Strategies for
microsatellite isolation: a review. Mol Ecol 11:1–16
Zeven AC (1967) The semi-wild oil palm and its industry in
Africa. Agric. Res. Rep. no. 698
Zulkifli Y, Maizura I, Rajinder S (2012) Evaluation of MPOB
oil palm germplasm (Elaeis guineensis) populations using
EST- SSR. J Oil Palm Res 24:1368–1377