Keywords

6.1 Introduction

Molecular markers can directly identify variation within organisms at the nucleic acid level, accurately distinguishing genetic variation from environmental variation. The discovery in the 1980s followed by the application of molecular technology in plants was a benchmark in breeding research in many of the important crop plants worldwide. While the initial applications of deoxyribonucleic acid (DNA) molecular markers were in annual plants, they also provided great potential to be used in perennial and challenging-to-breed plants including coconut. Conventional approaches of phenotypic evaluation of coconut germplasm and coconut breeding have resulted in achieving the objectives of such programmes to a considerable level. However, the use of molecular markers would be highly attractive and effective to overcome the inherent constraints of conventional approaches due to comparatively longer duration of these approaches in coconut. As is the case with many other crop plants, the molecular markers provide accurate and reliable approaches for evaluating the genetic diversity of coconut, as well as in the development of genetic linkage maps for marker-assisted selection (MAS) . The application of molecular markers in breeding research of coconut began in the 1990s. This chapter discusses the use of molecular markers in genetic diversity analysis, approaches to MAS made so far, and the future potential for wider applications in coconut.

6.2 Genetic Resources of Coconut

The coconut (Cocos nucifera L.) belongs to the monocotyledon family Arecaceae. The coconut genome is large representing 32 diploid chromosomes (2n = 2x = 32). The coconut palm does not record any closely related species or wild relatives; there are only different types of coconuts believed to be at different evolutionary and domestication stages.

6.2.1 Classification of Coconut Genetic Resources

Coconuts are broadly classified into two groups, Tall (typica) and Dwarf (nana) . There is a high phenotypic variation in coconut, and non-standard names are currently used to identify different coconut populations. This causes a lot of difficulties in developing a standard and formal systematic classification for coconut germplasm in the world. According to Menon and Pandalai (1958), the term “variety” denotes a single or a group of strains which differ from the other groups in structure or function and have the ability of true-to-type reproduction. However, using this definition, many reported types of coconuts cannot be considered as varieties. There are many different forms or types of coconut reported in the literature, but an acceptable and standard classification that describes the world coconut germplasm is yet to be formulated. Currently, different countries adopt different classifications to describe the genetic resources available in each part of the world.

The first classification of global coconut germplasm was recorded by Narayana and John (1949). In this classification, Tall and Dwarf coconuts were identified as distinct types of coconut. Both these types were divided into varieties based on the botanical features. Accordingly, Tall coconuts recorded three varieties, namely, typica, spicata, and androgena, while the Dwarf coconuts were divided into two varieties: nana and javanica. Of the Tall and Dwarf types, Tall coconut was the most abundant type throughout the world and the type that was planted on a commercial scale. In contrast, the Dwarf coconut type was less common and not grown on commercial scale. The variety nana of the Dwarf type was less vigorous but was reported to be early bearing reaching the reproductive phase approximately 3 years after planting, and the second variety javanica was more vigorous and comparatively late bearing taking about 5–6 years for floral initiation.

Based on a literature survey conducted by Gangolly et al. (1957), a coconut classification was presented by Menon and Pandalai (1958). This classification was like the previous classification by Narayana and John in identifying the two main types of coconut as Tall and Dwarf . In the same year, Liyanage (1958) reported a classification for Sri Lankan coconut germplasm, proposing three varieties, namely, typica (Tall), nana (Dwarf), and a new variety aurantiaca. The new coconut variety aurantiaca included coconuts of intermediate height between the Tall and Dwarf types (Table 6.1).

Table 6.1 General characteristics defining Tall, Dwarf , and Intermediate types of coconut
Full size table

A subsequent coconut classification was reported based on the pollination or breeding behaviour of coconut (Fremond et al. 1966). As per this classification, Tall and Dwarf coconuts , which were previously categorized as types, were named as two distinct coconut varieties: Tall characterized by cross-pollinating breeding behaviour, resulting in allogamous populations, and the Dwarf by self-pollinating breeding behaviour and autogamous populations. However, the Tall coconut variety is known to display low levels of self-pollination via inter-spadix pollination during favourable time periods for inflorescence emission. In addition, the Dwarf coconut variety naturally displays some degrees of cross-pollination when they are present in the vicinity of Tall coconuts (Whitehead 1976).

In 1978, Harries proposed a different classification, identifying two types of coconuts as “Niu kafa type ” and “Niu vai type”. The fruits of Niu kafa type were large with a thick husk and a low volume of water. They were long and angular in shape and were slow to germinate. In contrast, the fruits of ‘Niu vai type’ were more spherical shaped with a higher volume of endosperm and less husk content. They also were comparatively early germinating and were recorded as the more domesticated type of coconut that were selected and cultivated. Harries (1978) further suggested that the widely spread pan tropical coconut populations of today resulted from the introgression of the Niu kafa- and Niu vai-type coconuts and the selection and dissemination of them by humans (Harries 1978).

The identification of two main types of coconut is a common feature of the majority of coconut classifications to date. Each of these two groups displays certain phenotypic variations within the main type. In Tall coconuts, the size and shape of the fruit and shell and kernel thickness contribute mostly to this variation, Sri Lanka Tall Bodiri and Laccadive Micro Tall being examples. For Dwarf coconuts, the main feature contributing to within-type variation is the fruit colour which is indicated in the standard nomenclature as Sri Lanka Yellow Dwarf, Chowghat Green Dwarf, Cameroon Red Dwarf, etc. Such phenotypic variations within the type were referred to as forms within each variety by Liyanage (1958) and as phenotypically distinct variants by Bourdeix et al. (2005).

World coconut classifications do not include some of the different phenotypes present in certain countries and are showing semi-Tall or semi-Dwarf stature, Gangabondam in India and Niu Leka Dwarf in Fiji being examples. Therefore, there is an important need for an extensive evaluation and characterization of different types/forms of coconut grown worldwide in order to develop a standardized classification. A comprehensive classification should include a standard international nomenclature, local or vernacular names, and important specific descriptors for each coconut accession at the entry of them into a genebank for better utilization of coconut germplasm in breeding programmes.

6.2.2 Domestication of Coconut

Gunn et al. (2011) in the most extensive investigation to date reported the domestication history and the population structure of global coconut germplasm by analysing 1,322 palms spanning the geographical and phenotypic diversity of coconut at 10 microsatellite marker loci. The team reported two genetically distinct subpopulations relating them to Pacific and Indo-Atlantic Ocean basins and, accordingly, suggested two independent origins in the two regions; the islands in Southeast Asia and southern regions of the Indian subcontinent with the same population structure maintained to date despite the long-term mediation by humans in cultivation and dispersion. The selection by humans favoured the Niu vai type: more round-shaped nuts described by Harries (1978), Dwarf growth habit, and self-pollinating breeding behaviour. Gunn et al. (2011) located coconuts with admixture between the Pacific group and the Indo-Atlantic group in the south-western Indian Ocean and further revealed substructure within the Pacific coconuts. Further, data reveals that the original coconut gene pools in the South Asia and Atlantic oceans consist only of Tall-type coconuts. In contrast to the origin and domestication of Tall-type coconuts, the Dwarf coconuts have been found to be evolved from the cross-pollinating Tall coconuts in Southeast Asia. Dwarf coconuts have acquired self-pollinating breeding behaviour, but evolution from the cross-pollinating Talls and subsequent mutations have resulted in a considerable level of distinguishable phenotypic diversity in Dwarf coconuts (Coppens et al. 2018).

A study by Gunn et al. (2018) reports the distribution of coconut dating back to 28–44 million years ago, during the Eocene to Oligocene era, but that the distribution was limited to Southeast Asia and Oceania. The South Indian subcontinent coconuts came into being only 10,000 years ago when agriculture started, followed by domestication through cultivation. Based on archaeological studies, Summerhayes (2018) stated that the evidence from DNA data on the domestication of coconut may have the omission of ‘equifinality theory’ which states that similar results can be arrived at by two different processes. The author reveals that the coconuts in South Asia date back only to the mid-Holocene era, i.e. about 6,000 years ago, while coconuts were reported to be present in the Western Pacific region 20,000 years ago. Accordingly, combining DNA evidence with archaeological and palynological data, Summerhayes states that the coconut domestication has taken place in the Pacific followed by subsequent events of admixture elsewhere in the world.

6.3 Conservation of Coconut Genetic Resources

Globally, coconut genetic resources are endangered due to a relatively high rate of genetic erosion caused by several factors including urbanization and industrialization, infrastructure development, a shift from coconut cultivation to high-value cash crops, natural disasters and biotic stresses, and pests and diseases. In addition, the wide-scale replanting of coconuts from relatively few numbers of high-yielding improved cultivars, especially in areas of natural coconut stands where the genetic diversity is high, results in the dwindling of the genetic diversity of coconut germplasm. Consequently, a few decades ago, the identification, collection, and conservation of coconut genetic resources were recognized as an important objective in individual coconut-growing countries and later expanded to be a phenomenon of international interest, mainly because the future of coconut breeding would be based on the availability of diverse germplasm. Accordingly, in 1992, the international agency Coconut Genetic Resources Network (COGENT) was founded, under the umbrella of the International Plant Genetic Resources Institute (IPGRI , now renamed as Bioversity International). Strengthening and standardizing the individual country programmes and bringing the operations under an international forum for knowledge and material sharing were the main objectives of COGENT (http://www.cogentnetwork.org). In the year 1997, the COGENT initiated a systematic programme for collection and conservation of coconut genetic resources, which was funded by the Asian Development Bank.

6.3.1 Nomenclature of Accessions Conserved in Genebanks

In order to standardize and internationalize the coconut germplasm collection and conservation, the COGENT initiated the international Coconut Genetic Resources Database (CGRD) for the computerized cataloguing of the global coconut germplasm by facilitating the uploading of the data of national germplasm repositories by individual countries. Each conserved coconut accession, in the CGRD, is identified with a unique international name, comprising of the type (whether Tall or Dwarf), the geographical reference, and in the case of a Dwarf the respective colour form. West Coast Tall, Sri Lanka Tall, East Coast Tall, Malayan Yellow Dwarf, and Brazilian Green Dwarf are a few examples for this nomenclature in CGRD. In addition to this information, in the CGRD, accessions are assigned a more specific geographic location within each country in the naming of accessions conserved in the international field genebanks. Examples are San Ramon Tall in the Philippines and Sri Lanka Tall Ambakelle. Further studies will be essential to determine whether there is actual genetic variation among these “ecotypes” as they are referred to.

The standard data recording process of CGRD is aimed at enhancing the awareness and knowledge of global coconut germplasm for exchange of germplasm among countries for enriching the national coconut breeding programmes as required (Hamelin et al. 2005). The CGRD database facilitates the compilation of passport, characterization, and evaluation data for the accessions recorded in it. Each individual country was to provide the data on collected accessions, the passport data, and to proceed with the collection of characterization and evaluation data to be fed into the CGRD at each stage of characterization .

6.4 Evaluation of Coconut Genetic Diversity

The coconut germplasm collection and conservation programme resulted in the initiation of the ex situ field genebanks of coconuts as national germplasm repositories. Those countries which had conserved their limited germplasm collections used this opportunity to expand the collections and establish new ex situ field genebanks in a systematic manner. This programme further resulted in several international ex situ field genebanks, in addition to the national ex situ field genebanks of coconut. The value from conservation was to be the characterization of the conserved material to facilitate the utilization in coconut breeding programmes. However, to date, the conserved germplasm in these ex situ field genebanks has not been systematically characterized, apart from several scattered studies intended either to update the CGRD or to find desirable parental material for genetic improvement programmes indicating gaps in all aspects of characterization: morphological, molecular, and biochemical assays.

6.5 Morphological Assessment of Coconut Genetic Diversity

The assessment of genetic diversity of species was entirely dependent on the morphological assays during the early period of the nineteenth century. Scoring for morphological markers is less technology-demanding than molecular markers yet suffers from several drawbacks. The limited number of morphological markers, the manifestation of the environmental effects on the phenotype masking the genetic variation, and the growth stage-dependent expression of morphologies are some examples for such drawbacks of morphological markers. Despite the said disadvantages, morphological characterization, if properly designed to assess for the environmental effects, reveals the overall manifestation of the genetic potential of the plant, and accordingly, several studies have been carried out for assessing the genetic diversity of coconut by morphological means.

The availability of a comprehensive and well-defined descriptor list is a prerequisite for the process of morphological characterization. Bioversity International has coordinated the preparation of descriptor lists for important crop species in the world including coconut. These descriptor lists describe each accession of a crop in four basic data categories: passport, characterization, preliminary evaluation, and further characterization and evaluation.

Passport data are the most preliminary data to be collected of a germplasm recorded at the same time the collection is made and at the site of collection. This preliminary description is important for the identification and standard nomenclature of a germplasm accession. The site of collection (recorded as village/state/country), location in terms of longitude and latitude, collector’s number, date the collection was made, botanical and vernacular names, sample type (whether wild/weedy/landrace/cultivar, etc.), source (field/farm store/institute, etc.), and the site environmental characteristics are some examples for passport data parameters. Currently ethnobotanical information and digitalized location identifying, global positioning system (GPS), and the management of data via advanced databases have been introduced with the extensive adoption of novel technology.

Characterization data consists of the characters found to be highly heritable and can be easily distinguishable. These are qualitative traits which are stable across environments and are governed by one or a few major genes. Fruit shape and fruit colour are examples for such traits in coconut. During preliminary evaluation, a limited number of additional traits, which are useful to identify the specific germplasm, are scored. The scoring for evaluation data is complex because they involve the traits which are of quantitative nature and are influenced by the environment. Information on stem, leaf, inflorescence and flower, fruit, and seeds are the evaluation data of the germplasm. The use of adequate sample sizes and proper experimental designs are important criteria for accurate assessment of the genetic diversity for quantitative traits. Further characterization and evaluation of germplasm include the scoring for potential agronomic characters which are useful for crop improvement: physiology, pathology, entomology, cytogenetics, biochemistry, and recently molecular information. Internationally accepted norms are followed in scoring, coding, and recording of each of the descriptor states.

6.5.1 Descriptors for Morphological Characterization

Several research teams have published on the general morphology of the coconut palm in different geographic locations and reported a high phenotypic variation within as well as among populations for fruit traits. Accordingly, variations in fruit size, fruit shape, colour of the epicarp of the fruit, and proportional weight of various fruit components such as husk, shell, kernel (the solid endosperm), and water (the liquid endosperm variations) have been highlighted. A survey of the diversity of fruit components of coconuts in the South Pacific revealed a high diversity among populations displaying both wild-type characters and more domesticated traits and a range of mixed phenotypes of the two extremes (Ashburner et al. 1997a). The results of a study by Foale (1991) revealed different shapes, varying from angular to pear or near-spherical shapes (Figs. 6.1, 6.2, and 6.3) and from shorter to longer fruits of coconuts. Furthermore, higher morphological diversity has been observed in coconuts in Southeast Asia compared to those of South Asia, Africa, and South America (Benbadis 1992; Whitehead 1976).

Fig. 6.1
figure 1

Crown of a Tall-type coconut from South Asia having elongated fruit

Full size image
Fig. 6.2
figure 2

Crown of a Tall-type coconut palm from Southeast Asia and the Pacific region having large spherical-shaped fruit

Full size image
Fig. 6.3
figure 3

Crown of a Dwarf-type (yellow) coconut palm having smaller fruit

Full size image

Zizumbo-Villarreal (2005) studied the patterns of variations of coconut in Mexico using fruit morphological descriptors and reported three main Tall-type groups originating from different historical introductions. Several studies on the systematic characterization of coconut using standard descriptor lists have also been reported. Morphological diversity related to stem, leaf, inflorescence, and fruit morphology of Sri Lankan indigenous coconut varieties was studied (Perera and Ekanayake 2008), and the results revealed high levels of variation among the varieties examined. Further, Perera et al. (2009) characterized 24 germplasm accessions conserved in the ex situ field genebanks of Sri Lanka using morphological descriptors and recorded higher variations in Tall coconuts than in Dwarf accessions . Despite these attempts, comprehensive analysis of conserved coconut accessions for all the morphologies – stem, leaf, inflorescence, and fruit diversity – using morphological descriptors remains scarce due to various constraints and difficulties. However, attempts in the use of molecular markers for assessing the genetic diversity of coconut have been on the rise for the last two decades .

6.6 Molecular Approaches for Assessing Coconut Genetic Diversity

The use of molecular marker technology in coconut research was initiated in the mid-1990s and has continued to advance. Molecular markers have been applied in coconut for elucidating the genetic diversity of coconut germplasm, developing linkage maps, detection of intracellular pathogens, DNA finger printing of coconut, and hybridity testing. The initial attempts were on diversity analysis of coconut germplasm mainly with restriction fragment length polymorphism (RFLP), randomly amplified polymorphic DNA (RAPD), and amplified fragment length polymorphism (AFLP) markers, with the current trend of increased use of simple sequence repeats (SSR ).

6.6.1 Early Studies of Molecular Characterization of Coconut

One of the preliminary studies using molecular markers in assessing the genetic diversity of coconut has been reported in 1995. In this study, 17 varieties of coconut, representing different geographical regions, were evaluated through polymerase chain reaction (PCR) amplification of copia-like 16R1 repetitive elements (Rohde et al. 1995). The results of this study provided molecular evidence for the genetic similarity of Tall coconuts from East Africa to coconuts from the Indian Ocean and the grouping of the West Coastal Panama Tall coconuts with coconuts from the Pacific and the Southeast Asia.

In another preliminary research study, utilizing molecular markers in assessing the genetic diversity of coconut was the use of RFLP markers to study a pool of coconut palms representing different geographical zones (Lebrun et al. 1998). This research was carried out by hybridizing 9 cDNA clones and a mitochondrial DNA clone developed from rice and 1 ribosomal DNA clone derived from wheat with the DNA extracted from a coconut palm pool of 100 individuals of 10 Tall coconuts and 7 Dwarf coconuts. The results provided evidence for two groups of coconut representing the geographical regions: the first being the coconuts from Far East and the Pacific and the second representing coconuts in the South Indian subcontinent – India, Sri Lanka, and East Africa. In addition, a comparatively high level of diversity was observed in the coconuts from the Far East and the Pacific compared to the rest. Furthermore, the coconut variety Panama Tall clustered together with coconut varieties representing the Pacific region, while the African variety, West African Tall, grouped with the coconuts from South Asia.

In yet another earlier study, 17 coconut populations from the South Pacific region were subjected to RAPD analysis, and the results reported a within-population diversity of about 60% which is a considerably high value (Ashburner et al. 1997b). The results from this study further illustrated the occurrence of two geographic groups in addition to distinguishing two single populations in the 17 populations studied. The study thus revealed a low rate of gene migration among the coconut populations in the said region, with the possibility of founder effects, followed by selection during the coconut domestication in the Pacific. The team further recommended the germplasm collection programmes in the region focus on populations and not the individual palms to capture the high genetic variation that was observed among populations.

Other research teams reported diversity studies on coconut in respective local germplasm, using RAPD. Accordingly, the conserved coconut germplasm in Sri Lanka was analysed using RAPD markers, and the results indicated a rather narrow genetic base (Dasanayaka et al. 2003; Everard 1996). Duran et al. (1997) studied 48 East African Tall (EAT) coconut genotypes using 22 RAPD primers. A total of 238 amplicons were produced out of which 204 (86%) were polymorphic and others were monomorphic. The dendrogram constructed, based on the RAPD analysis of 48 coconut accessions, clustered into two major groups recording an average dissimilarity index of 0.35. Upadhyay et al. (2002) analysed 14 coconut accessions of indigenous and exotic origin, including Tall-, Dwarf-, and Intermediate-type coconuts, conserved ex situ in the Coconut Germplasm Centre in Kerala, India. The genetic relationships among the tested coconut accessions revealed that the Intermediate coconuts were genetically closer to Dwarf coconuts than the Tall coconuts.

Ratnambal et al. (2001) used RAPD markers to screen the marker polymorphism to identify a set of informative markers and to characterize coconut germplasm in India. Hundreds of primers were screened in this study to detect polymorphism and allelic diversity in coconut. Among these, only 34% primers were polymorphic recording 1–16 polymorphic bands per RAPD primer. Daher et al. (2002) assessed the genetic divergence among 19 coconut populations by RAPD. The markers used permitted the identification of each of the populations, showing that they were genetically different revealing the lack of duplicates in the collection. The genetic diversity of 30 ex situ conserved coconut accessions representing the Pacific and Nicobar Islands was genotyped with RAPD markers (Sankaran et al. 2012). The resultant dendrogram from cluster analysis revealed two main clusters and the distinct genetic variation among the accessions. The informativeness of the tested 13 RAPD primers was also revealed recording an average polymorphic information content (PIC) value of 0.29, with a range of PIC values from 0.46 to 0.17, for primers OPF-19 and OPH-25, respectively (Sankaran et al. 2012).

Jayalekshmy and SreeRangasamy (2002) reported that intervarietal variation could be detected by RAPD markers. Out of the ten primers used, seven gave amplification products showing polymorphism between Tall and Dwarf varieties. They also reported that RAPD markers appeared to be of high value for characterizing the genetic resources. The use of AFLP markers in the diversity analysis of coconut was reported by Perera et al. (1998) and Teulat et al. (2000), and during the same period, diversity studies using inverse sequence-tagged repeat (ISTR) markers were reported by Duran et al. (1997) and Rohde et al. (2000).

6.6.2 Development and Use of SSR Markers in Diversity Analysis of Coconut

The need for the use of coconut-specific molecular markers was later highlighted to expand and enhance the molecular marker applications in coconut. Accordingly, two groups (Perera et al. 1999; Rivera et al. 1999) successfully developed simple sequence repeat (SSR) or microsatellite markers in 1999, using Sri Lanka Tall and Tagnanan Tall, respectively, as the base genetic material. These SSR markers are co-dominant and thus have been immensely useful in the evaluation of the genetic diversity of coconuts. The use of SSR markers was highly advantageous to genotype coconuts, due to the highly heterozygous nature resulting from out-breeding pollination behaviour. Accordingly, the use of SSR markers was effective in providing clear information on genetic relationships and identification of representative collections (Dasanayaka et al. 2003; Perera et al. 2000, 2001, 2003; Meerow et al. 2003; Teulat et al. 2000), hybridity testing (Perera et al. 2004), identifying somaclonal variations in tissue-cultured coconuts, and developing linkage maps of coconut (Baudouin et al. 2006; Herran et al. 2000; Lebrun et al. 2001). Standardization of the techniques is a must for facilitating the comparison of results derived across laboratories in different countries. In order to fulfil this requirement, an SSR marker kit, comprising of markers at 14 microsatellite loci, was developed and relevant software for the analysis of data was introduced (Baudouin and Lebrun 2002).

The use of SSR markers has been the most common marker system used to date, for the evaluation of the coconut germplasm. The genetic diversity of coconut was evaluated at 8 SSR marker loci, by analysing 130 individuals representing 75 Tall and 55 Dwarf coconut ecotypes spanning the coconut-growing areas in the world (Perera 1999). The results revealed higher allele richness in Tall coconuts compared to Dwarf coconuts, thus indicating the comparatively low genetic diversity in the Dwarfs. Simultaneous studies including 20 coconut varieties from the southern coast of Asia and the Pacific and 31 varieties from the same region (Rivera et al. 1999; Teulat et al. 2000) reported results which agree with that of Perera (1999). Thirty-three Sri Lankan Tall coconut populations collected across the coconut-growing areas in the country were subjected to SSR analysis (Perera et al. 2001), and the results provided evidence for lack of population differentiation in Sri Lanka Tall coconuts.

Microsatellite marker analysis of a collection of 179 global coconut accessions, displayed comparatively higher genetic variation and a heterozygosity value of 30% in the naturally cross-pollinating Tall coconuts, compared to low genetic variation and a low heterozygosity value of 2.5% in the naturally in-breeding Dwarf coconuts (Perera et al. 2000, 2001, 2003). Further studies revealed very low (5%) population differentiation in Tall coconuts from Sri Lanka (Perera et al. 2001), contrary to findings by Ashburner et al. (1997b), who reported higher genetic diversity between populations in Southeast Asia and Pacific coconut germplasm. Both studies reiterate the need to adjust conservation strategies accordingly. These findings have assisted in changing the strategies in the collection of genetic resources in different parts of the world and in addition provided insights into the genetic base of coconut. As an example, the revealing of the narrow genetic base of Sri Lankan coconuts led to the revision of breeding programmes by incorporating imported exotic material as parents in recent crossing projects. Molecular analysis of global coconut germplasm, represented by 51 Tall accessions and 49 Dwarf accessions, revealed the presence of 2 major groups of Tall coconuts: the first being the Tall coconuts from Southeast Asia and the Pacific including Panama West Coastal Region Talls and the second group comprising of Tall coconuts from South Asia and East and West Africa. The entirety of the global Dwarf coconut germplasm grouped into a sub-cluster within the main cluster of coconuts from the Southeast Asia and the Pacific (Perera et al. 2003).

Information derived from microsatellite markers on the genetic variation in coconut germplasm agree with the findings of other molecular marker techniques ISTR (Rohde et al. 1995) and RFLP (Lebrun et al. 1998). The combined results of several of these studies revealed the presence of two main groups of coconut: the first being the Southeast Asian and the Pacific island coconuts and the second being the Indo-Atlantic coconuts. As per the molecular evidence, derived from the studies described so far, Dwarf coconuts in the world are classified as a subgroup within the main group of Tall coconuts from the Southeast Asia and the Pacific. In addition, a reduction of allelic diversity is revealed among the Dwarf coconut accessions, indicating the evolution of the Dwarf group from within the Tall coconuts from the Southeast Asia and the Pacific, supporting the findings of Teulat et al. (2000), indicating a common origin for global Dwarf coconut germplasm.

Diversity of the coconut chloroplast genome has been studied by the polymerase chain reaction (PCR) amplification of chloroplast DNA of a sample of 130 individual palms of global coconut germplasm, with the objective of identifying genetic lineages in coconut (Perera 1999, 2002). These studies revealed the lack of chloroplast variation of the tested samples, providing evidence for close ancestry within the tested coconut germplasm. Later, moving towards high-throughput genotyping of coconut germplasm, the high-throughput marker system Diversity Arrays Technology (DArT) was validated for coconut, elucidating the genetic diversity of a collection of Sri Lankan coconut germplasm (Perera and Kilian 2008).

6.6.3 Use of Molecular Markers to Screen National-Level Repositories of Coconut Germplasm

Since early studies on the development of basic theories and preliminary screenings, many countries and different research groups have attempted diversity analysis of coconut using molecular markers in their respective genebanks/countries for different purposes. The findings of such investigations are being used for better understanding of the true genetic diversity among the genetic resources conserved in in situ and ex situ field genebanks. Below are some examples of such studies.

Simple sequence repeat markers have been used to study the tolerance/susceptibility of coconut accessions of West African Tall, to lethal yellowing (Konan et al. 2007a, b). The analysis of data at 12 SSR marker loci revealed the clustering patterns of the susceptible accessions from that of the tolerant accessions. The results further indicated that the two groups were genetically distant from each other, as determined by the specific alleles and the frequency variation of shared alleles of the accessions. The same genotypic profiles were used to determine the genetic diversity of the accessions using the 58 alleles that resulted recording an average of 4.8 alleles per SSR marker locus. This study demonstrated the feasibility of large-scale molecular screening of coconut accessions for desirable traits, in selecting of parents for breeding programmes, and for developing mapping populations for tagging genes for lethal yellowing.

A total of 30 coconut samples comprising of 12 accessions from China and 18 accessions from Southeast Asia were evaluated using 30 new microsatellite markers developed with the sequence data generated from an Illumina transcriptome profile (Yong et al. 2013). The results displayed variable levels of allelic polymorphism among the accessions. The analysis of population structure with the same genotypic profiles revealed the Chinese accessions to be a subset of the Southeast Asian coconut accessions. Accordingly, the authors suggested that the evolution of the Chinese accessions had not been independent from that of the tested accessions in Southeast Asia. With the combined results of the population structure analysis and historical evidence, it was concluded that the dissemination of coconuts to the Hainan Province of China occurred along the sea currents and human-mediated dispersal was responsible for coconuts moving from Southeast Asia to the Yunnan Province in China.

Investigations were carried out to elucidate the genetic diversity of coconut accessions using the PCR-based molecular marker systems RAPD , inter-simple sequence repeat (ISSR), and SSR (Kandoliya et al. 2018), with a total of 45 markers representing 15 markers each from the above marker systems. The three marker systems, RAPD, ISSR, and SSR, generated 82, 82, and 28 bands, respectively. All the marker systems recorded higher percentages of polymorphic alleles while recording a few unique bands as well. In addition, the values of similarity coefficient of clusters of the marker systems recorded values ranging from 22 to 83% for RAPD, 26 to 86% for ISSR, and 50 to 97% for SSR. Accordingly, it was concluded that the performance of the three molecular marker systems was comparable and equally reliable in assessing the genetic diversity of coconut genotypes.

The SSR marker technology was used to evaluate the genetic diversity among 48 individual coconut palms collected from the lowland coastal belt in Kenya (Oyoo et al. 2016). The information derived from the 15 SSR marker loci, analysed with Popgene version 1.31, revealed genetic diversities ranging from 0.0408 for marker locus CAC68 to 0.4861 for marker locus CAC23, recording a mean of 0.2839. The investigation further revealed the marker loci which are polymorphic for Kenyan germplasm and more importantly revealed a high within-population variation of 28%, compared to a low variation of only 2% recorded between populations, suggesting the non-dependence of molecular variation in the region they are cultivated.

The genetic diversity of 14 coconut accessions was evaluated at 8 SSR marker loci (Pradeepkumar et al. 2011). The accessions grouped into three clusters in the dendrogram. Cluster 1 consisted of five accessions from New Guinea, Cluster 2 of four from French Polynesia, and Cluster 3 of five accessions from the South Pacific, giving evidence for the geographical variations. The effect of controlled pollination on the maintenance of the levels of genetic diversity upon was investigated in three Tall coconut accessions, namely, Mozambique Tall (MZT), Gazelle Peninsula Tall (GPT), and Tahitian Tall (THT) (Yao et al. 2013). The genotypic analysis of the parents (G0) and the progeny derived via controlled pollination (G1) at 15 SSR loci revealed a slight reduction of gene diversity, varying from 0.69 to 0.587, low values of Jaccard dissimilarity index varying from 0.072 to 0.133, and low levels of genetic diversity ranging from 0.005 to 0.007 between the parental and regenerated populations. Accordingly, it was concluded that the genetic integrity of the original accessions conserved in field genebanks can be maintained satisfactorily by controlled pollination in the rejuvenation process.

The genetic diversity within and between populations of Brazilian Tall coconuts, as represented by 195 palms belonging to 10 populations, was studied at 13 SSR marker loci (Ribeiro et al. 2010). The results revealed 68 alleles, averaging 5.23 alleles across populations, varying from 2 to 13 per locus and mean gene diversity (He) and observed heterozygosity (Ho) values of 0.459 and 0.443, respectively. The among population genetic distances varied from 0.034 to 0.390, and the results provided molecular evidence for the presence of two groups, the first comprising of the Baía Formosa, Georgino Avelino, and São José do Mipibu populations and the second consisting of the Japoatã, Pacatuba, and Praia do Forte populations. The comprehensive analysis of data indicated spatial genetic structuring of populations, by geographically close populations displaying higher genetic similarities.

SSR markers were utilized to evaluate the genetic variation of coconuts in the Andaman and Nicobar Islands of India (Rajesh et al. 2008). The results of this study revealed 7.35 alleles and an average heterozygosity of 0.29. A mean fixation index (FST) of 0.49 indicated a high level of population differentiation among the tested coconut accessions. The highest heterozygosity was observed in Tall coconut accessions, as expected, and most of the rare alleles were recorded in Tall coconuts sampled in the Nicobar Islands. Tall coconut ecotypes were reported to display greater heterozygosity values ranging from 0.18 to 0.37 in comparison with Dwarf coconuts, the values for which ranged from 0.03 to 0.07 (Thomas et al. 2013). In this study, a total of 90 sample trees, originating from 6 ecotypes, were evaluated, at 14 SSR marker loci. The differences in observed and expected heterozygosity in Tall ecotypes were an indication of the genetic basis of resistance to diseases, by the combined analysis of SSR marker data with the morphological data scored.

Studies have been conducted to determine the genetic variability among Tall coconut accessions conserved at the International Coconut Genebank for Latin America and the Caribbean using SSR markers (Loiola et al. 2016). The study revealed information for decision-making regarding the conservation of coconut germplasm and for the higher accuracy of selecting diverse parents to be utilized in crossing programmes, including selecting varieties for resistance to lethal yellowing. The molecular genetic diversity among 14 coconut accessions from India was determined at 8 SSR marker loci (Pradeepkumar et al. 2011). The eight markers produced a high level of polymorphism, with an average of 4.166 alleles per locus. The dendrogram developed, based on the SSR analysis, separated the 14 coconut accessions into 3 major clusters. The smaller similarity coefficient value indicated the absence of similarity between the genotypes.

Rajesh et al. (2012) studied the genetic purity among coconut hybrids at 50 SSR marker loci. Chowghat Green Dwarf (CGD) and West Coast Tall (WCT) were used as parents for hybrid production. Among 50 SSR markers, 17 displayed the complementary banding patterns of both the parents, and the selfed progenies showed the banding patterns of only the mother palm. The study also revealed the importance of SSR markers in hybridity testing. Kriswiyanti et al. (2013) determined the genetic variation of Tall coconuts in Bali, based on the analysis at six SSR marker loci. In total, 80 alleles were identified with an average of 13.33 alleles per locus. The mean values of gene diversity and observed heterozygosity were 0.883 and 0.542, respectively. Gene diversity ranged from 0.85 to 0.92, with a mean of 0.88; the overall results explained a high genetic diversity among the Tall coconuts in Bali.

6.6.4 Expansion of Molecular Diversity Studies for Specific Findings

The early attempts on the use of molecular markers for diversity analysis only served the intended purpose. Subsequently, the molecular marker research was planned in such a way as to reveal additional information on the studied genetic resources. Accordingly, Kamaral et al. (2014) conducted a research study to characterize 15 Sri Lanka Yellow Dwarf (SLYD) coconut palms using 10 SSR markers. All 10 microsatellite primers produced polymorphic amplicons resulting in 34 alleles, scored in the 15 individuals of SLYD palms. A total of 22 heterozygous loci were identified with the results further revealing the existence of high genetic diversity within the SLYD coconut palms.

Shalini et al. (2007) conducted a study using 3 coconut populations with varied yield traits (high, medium, and low) at 32 SSR marker loci to determine the genetic variation. High- and medium-yielding populations showed maximum heterozygosity, which indicated that they are not undergoing a population expansion. But the low-yielding population exhibited a significant deficiency in genetic diversity, which indicated that they are undergoing a rapid population expansion. Further advancing the genetic diversity studies, the molecular marker technology was adopted for the determination of the population structure of coconut. The studies on determining the population structure and genetic diversity revealed relevant genetic information on Florida coconuts, with special reference to Fiji Dwarf coconut cultivars (Meerow et al. 2003).

Gunn et al. (2011) reported one of the most comprehensive studies on the domestication of coconut through the analysis of the population structure of 1,322 accessions, representing wide geographical and phenotypic diversity, using 10 microsatellite markers. The study concluded independent origins and persistent population structure of coconuts in the two main regions, despite the long-term cultivation of coconut and human-mediated dispersal. Genetic diversity and the population structure were studied in Sri Lankan Yellow Dwarf coconut phenotypes by Kamaral et al. (2016). In this experiment, the Yellow Dwarf coconut variety was purified from a mixture of phenotypes, and a novel semi-Tall self-pollinating coconut phenotype, termed Sri Lanka Yellow Semi Tall, was identified.

6.7 Genetic Linkage Mapping in Coconut

In addition to being used extensively in genetic diversity studies, molecular markers have been used in linkage and QTL mapping of coconut. Selection of or construction of a suitable mapping population is a crucial step in linkage mapping, while it is a must to have a dense coverage of molecular markers. A population for linkage mapping should be segregating for traits to be mapped. Basic segregating populations, such as the early filial generations and backcrosses or the advanced segregating populations including recombinant inbred lines (RILs) or doubled haploid lines (DHLs) , are being used for linkage mapping for self-pollinated crops. The development of suitable mapping populations is a highly challenging task in coconut, hindering the process and resulting in linkage maps lacking the power for fine location of gene/QTL.

Accordingly, the information generation in QTL mapping in an outbred species, including coconut, is concentrated on the available pedigree populations. However, in coconut, the family sizes are limited in numbers, preventing the formation of a sufficiently large mapping population. Formation of a large number of families, followed by the analysis of data to model the genic inheritance of multiple pedigree populations with appropriate statistical software, would be a solution to this problem (Kearsey and Luo 2003). The naturally outbreeding nature of Tall coconuts, resulting in heterozygous individuals, and the inherently inbreeding characteristic of Dwarf coconuts are a specific feature in coconut, facilitating the development of a segregating population by crossing the Tall with Dwarf coconuts (Bandaranayake 2006). Yet, for the resulting population to possess enough levels of segregation, the selected Tall parent should possess heterozygosity and polymorphism at important loci (Perera 2010).

The success of fruit setting upon artificial hand pollination of coconut is generally low, resulting in low numbers of seeds and progeny from a single mother tree. Due to this, it takes a considerable period to produce a population of reasonable size, inducing a long age gap among the progeny. This limitation can be overcome by the combination of progeny from several half sib families of a short age gap, in a single mapping population. A practical approach to develop such a mapping population would be to select a highly heterozygous male parent and pollinate enough Dwarf female parents using the pollen of the selected Tall male parent (Perera 2010).

6.7.1 Linkage and QTL Maps of Coconut for Marker-Assisted Selection

Despite the above-mentioned difficulties , several research groups have developed linkage maps for coconut. The first genome map of coconut was constructed with an F1 population of a cross between East African Tall and Laguna Tall (Rohde et al. 1999). The genotypic data for this map was derived with ISTR markers. A second genome map of coconut was produced in the Philippines, using a mapping population developed by crossing Malayan Yellow Dwarf with Laguna Tall. The molecular markers – AFLP, ISTR, RAPD, and ISSR – were utilized to derive the genotypic data for this linkage framework map, which positioned 382 makers covering the 16 linkage groups of coconut. The QTL map of the second mapping population was successful in identifying six QTLs governing early germination (Herran et al. 2000), providing the opportunity for marker-assisted selection in coconut. The second QTL map was expanded to include QTL governing vegetative traits leaf production and girth (Ritter et al. 2000). A subsequent mapping population constructed in the Ivory Coast, using Cameroon Red Dwarf and Rennell Island Tall as parents, resulted in a framework map anchoring 280 markers, in addition to identifying QTL for yield traits: numbers of nuts, bunches , and fruit components (Baudouin et al. 2006; Lebrun et al. 2001).

6.7.2 Improvements for Genetic Linkage Mapping in Coconut

Any mapping population should be developed using phenotypically and genotypically segregating parents, and the population itself should comprise of enough individuals. In coconut, a simulation study revealed the optimum size of a linkage mapping population to be about 400 individuals.

Evaluation of global coconut germplasm has resulted in the identification of two main groups of coconut: Southeast Asia and the Pacific group being the first and the South Asian and Atlantic group being the second. Accordingly, it is expected that maximum segregation would be the result in a cross between these two groups, making a highly informative population for linkage mapping. However, the earlier-mentioned mapping populations were derived from the crossing of varieties included in the same group, resulting in uninformative non-segregating loci in mapping. Given that about 84% of DNA loci generated in the Malayan Yellow Dwarf and Laguna Tall mapping population is monomorphic provides an example, revealing identical alleles at many loci between parents.

The availability of polymorphic markers for a mapping population is essential for gene mapping in coconut for a dense marker coverage to saturate the 16 linkage groups of coconut. Currently this requirement can be fulfilled with the availability of a large collection of SSR markers and the possibility to move forward with high-throughput marker systems.

6.8 Conclusions

The availability of many polymorphic markers and the facilities for high-throughput genotyping and sequencing enables the creation of accurate and reliable information from research on diversity analysis and on population structure. The density of molecular marker coverage of the coconut genome is a crucial consideration in fine-scale characterization and evaluation of germplasm and genome mapping in coconut. The molecular marker kit developed in 2002 for the analysis of coconut comprises of 14 genomic SSR marker loci (Perera et al. 2018; Pokou et al. 2018). However, with the development of sequencing projects, which have become faster and more economical, research needs to be directed towards identifying a comprehensive set of more targeted loci representing the coconut genome. The representative marker loci should cover functional loci of important phenotypic and agronomic traits for molecular marker studies to be more effective and practically useful. Here again concerted efforts of the coconut molecular biologist are needed to decide the molecular marker system or combined systems to be used to develop high-throughput systems, which are affordable and feasible, to be used even by resource-poor laboratories in coconut-growing countries. However, such research should be coordinated at an international level to develop standard methodologies for the molecular characterization of coconut. The information thus derived will be more targeted and useful for formulating and refining further collection and conservation of coconut germplasm, management of genebanks, identification of duplicates, and determining the strategies for rejuvenation of the existing field genebanks. It could also enhance the utilization of genebank material by assisting in parental selection in coconut breeding programmes aimed at combining the desirable characters from diverse parents into novel cultivars. Such measures will bring the much-awaited benefits of molecular marker techniques to farmers and other stakeholders of the coconut value chain.