**5. Utility of phylogenetic tree in revealing genetic diversity**

In the assessment of genetic diversity it is important to define relationships existing within or among sets of germplasm collection of a species and their evolutionary history within or with related species. Phylogenetic trees serve as extremely powerful tools for organizing and illustrating these relationships. Phylogenetic trees have been successfully used in guiding conservation and biodiversity efforts (Sul et al., 2009) and establish relationship of cacao with its wild relatives (Figuera et al., 1994). A phylogenetic tree is a diagrammatic branching "tree" illustrating evolutionary relationships among entities within a species or various biological species based on similarities and differences in their physical and/or genetic characteristics. Organisms with similar morphologies or DNA sequences are likely to be more closely related than organisms with different structures or sequences. Entities that are joined together in the tree are implied to have evolved from a common ancestor. Each branch point represents the divergence of two species while sister taxa are groups that share an immediate common ancestor. Phylogenetic trees can either be rooted or unrooted. A rooted phylogenetic tree clearly shows relationship of each entity with the (usually imputed) most recent common ancestor of all the entities. Rooted trees are often constructed with the use of a definitive related 'outgroup' taxa. An 'outgroup' is a species or group of species that is closely related to the 'ingroup', the various species being studied. Unrooted trees, on the other hand, depict the relatedness of the entities without making assumptions about their ancestry. The principles of maximum parsimony and maximum likelihood are often used to analyze phylogenetic relationships with computer programs. The principle of maximum parsimony assumes that the tree that requires the fewest evolutionary events (appearances of shared derived characters) is the most likely. The principle of maximum likelihood states that, given certain rules about how DNA changes over time, a tree can be found that reflects the most likely sequence of evolutionary events.

#### **6. Genetic diversity analysis results from West and Central Africa**

Prior to recent studies in Cameroon, Cote d'Ivoire, Ghana and Nigeria, there was no useful information on the extent of genetic diversity in the cocoa cultivated in West and Central African countries. These studies were conducted within the framework of the Sustainable Tree Crops Program, a public-private partnership platform endowed by the United States Agency for International Development, US Department of Agriculture and chocolate industry partners such as the Mars Incorporated. These studies assessed genetic diversity in the introduced primary cacao clones and germplasm accessions used to develop improved hybrids distributed to farmers and cacao accessions on farmers' fields across the sub-region. Microsatellite markers were used to assess genetic diversity in these accessions. In addition to the microsatellite studies, some studies were also carried out to determine variation in agro-morphological and phenotypic characteristics of cacao germplasm in farmers and genebank collections.

#### **6.1 Cameroon (Efombagn et al., 2006; Efombagn et al., 2008; Efombagn et al. 2009)**

In their study of some 194 cocoa accessions collected in farms in Southern Cameroon during field surveys (Plate 1 d) and 71 Trinitario and Upper Amazon clones available in genebank collections on-station were assessed using 13 SSR markers. The gene diversity, genetic differentiation and genetic similarities were analyzed for the different populations. In total, 282 alleles were detected within all the populations studied (Plate 1c). The farm accessions were strongly differentiated based on their geographical origin, with accessions coming from the East province clustering together with local Trinitario accessions from the

with related species. Phylogenetic trees serve as extremely powerful tools for organizing and illustrating these relationships. Phylogenetic trees have been successfully used in guiding conservation and biodiversity efforts (Sul et al., 2009) and establish relationship of cacao with its wild relatives (Figuera et al., 1994). A phylogenetic tree is a diagrammatic branching "tree" illustrating evolutionary relationships among entities within a species or various biological species based on similarities and differences in their physical and/or genetic characteristics. Organisms with similar morphologies or DNA sequences are likely to be more closely related than organisms with different structures or sequences. Entities that are joined together in the tree are implied to have evolved from a common ancestor. Each branch point represents the divergence of two species while sister taxa are groups that share an immediate common ancestor. Phylogenetic trees can either be rooted or unrooted. A rooted phylogenetic tree clearly shows relationship of each entity with the (usually imputed) most recent common ancestor of all the entities. Rooted trees are often constructed with the use of a definitive related 'outgroup' taxa. An 'outgroup' is a species or group of species that is closely related to the 'ingroup', the various species being studied. Unrooted trees, on the other hand, depict the relatedness of the entities without making assumptions about their ancestry. The principles of maximum parsimony and maximum likelihood are often used to analyze phylogenetic relationships with computer programs. The principle of maximum parsimony assumes that the tree that requires the fewest evolutionary events (appearances of shared derived characters) is the most likely. The principle of maximum likelihood states that, given certain rules about how DNA changes over time, a tree can be

found that reflects the most likely sequence of evolutionary events.

genebank collections.

**6. Genetic diversity analysis results from West and Central Africa** 

Prior to recent studies in Cameroon, Cote d'Ivoire, Ghana and Nigeria, there was no useful information on the extent of genetic diversity in the cocoa cultivated in West and Central African countries. These studies were conducted within the framework of the Sustainable Tree Crops Program, a public-private partnership platform endowed by the United States Agency for International Development, US Department of Agriculture and chocolate industry partners such as the Mars Incorporated. These studies assessed genetic diversity in the introduced primary cacao clones and germplasm accessions used to develop improved hybrids distributed to farmers and cacao accessions on farmers' fields across the sub-region. Microsatellite markers were used to assess genetic diversity in these accessions. In addition to the microsatellite studies, some studies were also carried out to determine variation in agro-morphological and phenotypic characteristics of cacao germplasm in farmers and

**6.1 Cameroon (Efombagn et al., 2006; Efombagn et al., 2008; Efombagn et al. 2009)** 

In their study of some 194 cocoa accessions collected in farms in Southern Cameroon during field surveys (Plate 1 d) and 71 Trinitario and Upper Amazon clones available in genebank collections on-station were assessed using 13 SSR markers. The gene diversity, genetic differentiation and genetic similarities were analyzed for the different populations. In total, 282 alleles were detected within all the populations studied (Plate 1c). The farm accessions were strongly differentiated based on their geographical origin, with accessions coming from the East province clustering together with local Trinitario accessions from the genebank while accessions from the Centre-South provinces clustered with Amazon and hybrid accessions (Plate 1b), suggesting greater uptake of seed garden materials in farms in these provinces. The genetic diversity parameters indicate that the farmers' planting material was not highly diverse, but genetically close to parental genotypes available in genebanks (Plate 1a and 1c). However, some promising Upper Amazon clones (T-clones) that have also been used as parents of released hybrid varieties were genetically distant from the accessions. Their result suggested that the progenies of the Upper Amazon parents have so far been poorly used in the cocoa farms surveyed. A large genetic diversity was observed in the farm (*Hnb* = 0.34 – 0.72) and genebank (*Hnb* = 0.64 – 0.66) materials (Plate 1c). The large variability observed in farmer plantations was attributed to the large variation of first cocoa introductions (Bartley 2005), and the introduction of UA germplasm in the 1950's with its subsequent use in the cocoa breeding program. They also observed a higher private allelic richness in farm genotypes (*A*p= 2.03) than those of SNK (Selection of Nkoemvone), ICS (Imperial College Selection), T (Trinidad) and UPA (Upper Amazon) clones. This indicated that farm accessions also harbor some genes that are not present in current national field genebanks. Evidence were found of admixture in farmers' fields which must have been due to hybridization (in seed gardens) and to substantial natural recombination in farmers' fields. Also, since farmers tend to use seeds issued from open pollination in their plantations for new plantings (replacement of dead cacao trees, extension or the creation of new cacao farms), these must have resulted in the presence of admixture as observed in these materials. Surprisingly however, was the observation that there was no relationships between the 'Criollo' reference materials with cacao accessions collected from farmers plantations (Plate 1a).

In a study to determine morphological diversity existing in cacao farms in relation to genetic diversity in gene bank accessions a total of 300 farm accessions (FA) were selected in the two major cocoa producing areas (Southern and Western) of Cameroon. Seventeen quantitative and qualitative descriptors used in this study were related to leaf (flush colour), flower (ligule colour), pod (weight, length, width, apex form, shape, rugosity, colour, husk hardness, basal constriction and pod index) and seed (number, length, width, dry weight and colour) characters. For the qualitative characters evaluated, considerable morphological variation was observed using the Shannon Weaver diversity index (SWDI) within FA and gene bank accessions. Among the FA, a differentiation between southern and western regions was only possible when using quantitative pod traits. Mean quantitative traits values of FA were not too different than those of most gene bank AGs, except for a few traits of agronomical interest (seed weight and pod index). No significant variation was observed for seed traits in all FA groups (southern/western). The morphological structure (quantitative traits) showed spatial differentiation between western and southern FA and a closer relationship between gene bank and some farm accessions.

#### **6.2 Cote D'Ivoire (Pokou et al., 2009; Tahi et al., 2008)**

Since the introduction of the 'Amelonado' type in 1880 that was widely cultivated in cocoa growing regions (Plate 2a), the first step at genetic improvement of the locally available germplasm took place with mass selection of 'better' types in local farms between 1947 and 1958. Mostly due to low genetic variability available in local types as is the case in other West African countries, the Upper Amazon Forastero types were introduced in 1954. This

b)

a)

b)

Plate 1. a. A scatter plot showing genetic structure of planting materials in farmers and genebank materials in Cameroon; b. Spatial genetic differentiation between farmers accessions in central and south provinces from accessions in eastern province of Cameroon; c. genetic diversity indices of farmers and genebank accessions of cocoa in Cameroon; d. locations of cacao accessions collected within cocoa producing area of Cameroon used for the study (Source: Efombagn et al., 2006; Efombagn et al., 2009).

culminated in the development and distribution of selected hybrids between from the 1960s. However, much impact was made with the distribution on the hybrids developed and distributed in 1975 (Besse, 1975). In a survey conducted recently, it was shown that 71 % of materials grown on farms were locally selected 'Amelonado' type, 23 % selected improved hybrid types, while 6 % of the farms were grown with a mixture of local and improved types (Plate 2c).

culminated in the development and distribution of selected hybrids between from the 1960s. However, much impact was made with the distribution on the hybrids developed and distributed in 1975 (Besse, 1975). In a survey conducted recently, it was shown that 71 % of materials grown on farms were locally selected 'Amelonado' type, 23 % selected improved hybrid types, while 6 % of the farms were grown with a mixture of local and improved

types (Plate 2c).

a)

b)

Plate 2. a. Cocoa producing regions in the humid forest of Cote d'Ivoire; b. genetic diversity in cacao accessions in farmers fields in Abengourou (green vertical lines) in relation to seed garden materials (red vertical lines); c. genetic diversity in cacao accessions in farmers fields in Divo (green vertical lines) in relation to seed garden materials (red vertical lines); d. distribution of different cacao types present in farmers fields in Cote d'Ivoire.

In a study conducted between 2003 and 2005 (Pokou et al., 2009), 12 microsatellites (simple sequence repeats marker) were used to asses genetic diversity of cacao types. Results showed considerable diversity in farmers accessions reflecting largely hybridization between local Amelonado types and Upper Amazon types distributed in the 1970s (Plates 2b & 2c). However, a significant proportion of diversity in seed garden materials was yet to diffuse to the farms indicating that farmers still largely used their own 'selected' materials against those developed in research institutions. A further analysis was carried of the reciprocal recurrent selection programme set up in 1990. This involved two main genetic groups: Upper Amazon Forastero (UA) and a mixture of Lower Amazon Forastero (LA) and Trinitario (T). Based on data obtained from 12 microsatellite primers, the genetic diversity and genetic distances of the parental populations used in the first and second selection cycles are presented. The results revealed that the diversity of populations UA0 and UA1 on the one hand and (LA+T)0 and (LA+T)1 on the other is similar. The genetic distances were small between the parental populations used for the first and second cycles. Genetic diversity was greater in the UA group than in the LA+T group. The number of rare and of private alleles was reduced for both genetic groups, as well as the number of the frequent alleles in the LA+T group.

#### **6.3 Ghana (Opoku et al., 2007)**

In order to assess the genetic diversity of cacao types grown, some 377 accessions including farmers' accessions, breeders' collection and parental clones were collected from all cocoa growing regions of Ghana (Plate 3a), and analyzed using 17 microsatellite markers. Genetic diversity indices indicated that average gene diversity was high in all populations, with mean observed heterozygosity of 0.738 (Plate 3b). Although the highest was recorded in accessions from breeders' and parental collections, genetic diversity in the farmers' collection was comparatively high. Included in the study were a few extant trees among one of the earliest Tetteh Quarshie's introduction in the late 19th century.

b)

202 Genetic Diversity in Plants

against those developed in research institutions. A further analysis was carried of the reciprocal recurrent selection programme set up in 1990. This involved two main genetic groups: Upper Amazon Forastero (UA) and a mixture of Lower Amazon Forastero (LA) and Trinitario (T). Based on data obtained from 12 microsatellite primers, the genetic diversity and genetic distances of the parental populations used in the first and second selection cycles are presented. The results revealed that the diversity of populations UA0 and UA1 on the one hand and (LA+T)0 and (LA+T)1 on the other is similar. The genetic distances were small between the parental populations used for the first and second cycles. Genetic diversity was greater in the UA group than in the LA+T group. The number of rare and of private alleles was reduced for both genetic groups, as well as the number of the frequent

In order to assess the genetic diversity of cacao types grown, some 377 accessions including farmers' accessions, breeders' collection and parental clones were collected from all cocoa growing regions of Ghana (Plate 3a), and analyzed using 17 microsatellite markers. Genetic diversity indices indicated that average gene diversity was high in all populations, with mean observed heterozygosity of 0.738 (Plate 3b). Although the highest was recorded in accessions from breeders' and parental collections, genetic diversity in the farmers' collection was comparatively high. Included in the study were a few extant trees among one

of the earliest Tetteh Quarshie's introduction in the late 19th century.

a)

alleles in the LA+T group.

**6.3 Ghana (Opoku et al., 2007)** 


Plate 3. a. Cocoa regions of Ghana indicating sites of cacao accessions collection (dots); b. Table of genetic diversity indices in cacao form the different regions and germplasm collection; c. Relationships among cacao accessions in the different regions and field genebank accessions

The cocoa produced at the Tetteh Quarshie farm is of the Amelonado type, which originated from Brazil. Results of diversity analysis showed a clear separate clustering of accessions from Tetteh Quarshie farm and Aburi garden compared to rest of cocoa populations presently grown n Ghana (Plate 3c). These accessions were among the first introductions into the country and then spread across the country; however, the results indicated that presently these accessions have little or no influence on the current plantings in farmers' fields. Subsequent hybridization with later introductions and adoption of farmers' own or newly released improved germplasm might have been responsible for observation made. The accessions from Western region clustered separately from those of breeders' collection and populations of other regions indicating that the breeders' germplasm had less impact on planting materials in the West of the country in comparison with other regions. This proved to be in agreement with the historical records that cocoa cultivation in Ghana had spread from other adjacent regions to the West. Additionally, the seed gardens from which farmers could obtain improved planting materials developed by breeders are fewer in the region and are inaccessible due to poor road network. On the other hand, farmers from other regions mostly collected seeds from the 'Seed gardens'. This explained why the farmers' collections from these regions clustered with the Breeders' collections. Another interesting observation was that the germplasm from Central, Ashanti, Volta, and Eastern regions, which constituted the earliest cocoa-growing regions of Ghana, clustered together and separately from the Parental clones and Breeders' collection, whereas the accessions from Brong Ahafo region clustered with the Breeders' collections. This showed that most of Brong Ahafo plantings were done at the time when the "Series II hybrids" had been developed and were popular in the country. Most farmers in this region might have used those varieties as planting material. However, in the case of the other regions, a substantial number of farmers had, in addition to breeders' varieties, used materials from their own farms or other neighboring sources.

#### **6.4 Nigeria (Aikpokpodion 2007; Aikpokpodion et al., 2009; Aikpokpodion 2010; Aikpokpodion et al., 2010)**

In a recent study, 12 microsatellite markers were used to determine genetic diversity in 574 accessions representing eight groups covering parental populations in West Africa, genebank, and farmers' accessions collected from cocoa growing regions of Nigeria (Plate 4a). From this study, it was shown that appreciable genetic diversity was present in on-farm and field genebank collections. A total of 144 alleles were detected in these accessions with a mean allelic richness of 4.39 alleles/ locus. The largest genetic diversity was found in the Upper Amazon parent population (Hnb=0.730), followed by the 1944 Posnette's Introduction (Hnb=0.704), and was lowest in the Local parent population (Hnb=0.471). Gene diversity was appreciably high in the farmers' populations (Hnb=0.563– 0.624); however, the effective number of alleles was lower than that found in the genebank's Posnette population. Fixation index estimates indicated deficiency of heterozygotes in the Upper Amazon and the Local parent populations (Fis=0.209 and 0.160, respectively), and excess of heterozygotes in the Trinitario parent population (Fis=−0.341). The presence of inbreeding in the Local parent populations and substructure (Wahlund effect) in the Upper Amazon were suggested for the deficiency of heterozygotes observed. In Nigeria, restricted gene flow and spatial differentiation was evident in cacao varieties grown by farmers in Nigeria (Plate 4b). Cacao trees grown on farmers' fields in southwestern and mid-western Nigeria are mainly hybrids of the Upper Amazon and the local Amelonado varieties, while the local Amelonado variety predominates in southeastern Nigeria (Plate 4c).The non-significant genetic differentiation observed between the genebank's and farmers' populations indicated significant impact of national breeding programs on varieties grown in farmers' plantations (Plate 4c). Results also showed that a small proportion of the genetic diversity available in field gene banks at the Cocoa Research Institute of Nigeria (CRIN) had been used to develop improved varieties

fields. Subsequent hybridization with later introductions and adoption of farmers' own or newly released improved germplasm might have been responsible for observation made. The accessions from Western region clustered separately from those of breeders' collection and populations of other regions indicating that the breeders' germplasm had less impact on planting materials in the West of the country in comparison with other regions. This proved to be in agreement with the historical records that cocoa cultivation in Ghana had spread from other adjacent regions to the West. Additionally, the seed gardens from which farmers could obtain improved planting materials developed by breeders are fewer in the region and are inaccessible due to poor road network. On the other hand, farmers from other regions mostly collected seeds from the 'Seed gardens'. This explained why the farmers' collections from these regions clustered with the Breeders' collections. Another interesting observation was that the germplasm from Central, Ashanti, Volta, and Eastern regions, which constituted the earliest cocoa-growing regions of Ghana, clustered together and separately from the Parental clones and Breeders' collection, whereas the accessions from Brong Ahafo region clustered with the Breeders' collections. This showed that most of Brong Ahafo plantings were done at the time when the "Series II hybrids" had been developed and were popular in the country. Most farmers in this region might have used those varieties as planting material. However, in the case of the other regions, a substantial number of farmers had, in addition to breeders' varieties, used materials from their own farms or other

**6.4 Nigeria (Aikpokpodion 2007; Aikpokpodion et al., 2009; Aikpokpodion 2010;** 

In a recent study, 12 microsatellite markers were used to determine genetic diversity in 574 accessions representing eight groups covering parental populations in West Africa, genebank, and farmers' accessions collected from cocoa growing regions of Nigeria (Plate 4a). From this study, it was shown that appreciable genetic diversity was present in on-farm and field genebank collections. A total of 144 alleles were detected in these accessions with a mean allelic richness of 4.39 alleles/ locus. The largest genetic diversity was found in the Upper Amazon parent population (Hnb=0.730), followed by the 1944 Posnette's Introduction (Hnb=0.704), and was lowest in the Local parent population (Hnb=0.471). Gene diversity was appreciably high in the farmers' populations (Hnb=0.563– 0.624); however, the effective number of alleles was lower than that found in the genebank's Posnette population. Fixation index estimates indicated deficiency of heterozygotes in the Upper Amazon and the Local parent populations (Fis=0.209 and 0.160, respectively), and excess of heterozygotes in the Trinitario parent population (Fis=−0.341). The presence of inbreeding in the Local parent populations and substructure (Wahlund effect) in the Upper Amazon were suggested for the deficiency of heterozygotes observed. In Nigeria, restricted gene flow and spatial differentiation was evident in cacao varieties grown by farmers in Nigeria (Plate 4b). Cacao trees grown on farmers' fields in southwestern and mid-western Nigeria are mainly hybrids of the Upper Amazon and the local Amelonado varieties, while the local Amelonado variety predominates in southeastern Nigeria (Plate 4c).The non-significant genetic differentiation observed between the genebank's and farmers' populations indicated significant impact of national breeding programs on varieties grown in farmers' plantations (Plate 4c). Results also showed that a small proportion of the genetic diversity available in field gene banks at the Cocoa Research Institute of Nigeria (CRIN) had been used to develop improved varieties

neighboring sources.

**Aikpokpodion et al., 2010)** 

in the Institute's cacao breeding programs and that 'Scavina' and 'Imperial Mixed Calabacillo' cacao varieties of the Upper Amazon Forastero have not been significantly utilized in Nigerian cacao breeding programs. Population structure analysis cacao types grown in farmers' fields showed that the Upper Amazon Forastero constitute 66 % of cacao grown, Amelonado made up 24 %, Trinitario accounted for 6 % while other types made up 4 % of cacao grown (Plate 4d). Large number of alleles were found in the farmers' populations (7.67–9.00), and compared with the number recorded for the genebank's collection, although the effective number of alleles (4.49–4.80) was lower than in Posnette's population (5.65). This situation is encouraging as it indicated that genetic diversity held in farmers' collection in commercial plantations is much greater than, and showed a major shift from, what it was since introduction in the late nineteenth century till the 1950s when the highly uniform local Amelonado cocoa was predominant on the field. We can conclude from the molecular data used in this study that there has been a significant variety replacement of the Amelonado cacao grown on fields in West Africa with Upper Amazon and Amazon × Amelonado hybrids.

In a study to determine phenotypic variation among cacao grown in Nigeria, 17 agromorphological traits were studied in 184 accessions collected from farmers' fields (138) and field genebank collections (46). Fruit and bean traits of Upper Amazon Forasteros observed in farmers' accessions provided evidence of a shift from previously grown local 'West African Amelonado' from the Lower Amazon Forastero population. The large variation observed in this study for cacao grown by farmers indicated a high level of heterogeneity

Plate 4. a. Sites of cacao germplasm collection(red dots) in cocoa producing region of Nigeria; b. Relationships showing spatial differentiation among cacao accessions collected in farmers' fields in ideal climate (yellow), ideal soil (blue) and marginal climate (white) conditions in Nigeria; c. Relationships between farmers and field genebank cacao accessions in Nigeria; d. Population structure indicating cacao types grown on farmers fields in Nigeria.

in materials maintained on-farm. The low percentage of fruit traits that are typical of 'Amelonado' and 'Trinitario' types provides some evidence of variety replacement of 'West Africa Amelonado' (WAA) cacao types in farmers' fields with Upper Amazon-derived types. This showed a radical shift from the situation preceding the 1950s, when uniform 'Amelonado' cacao types were mainly grown. This would have resulted from the use Upper Amazon-derived cacao varieties distributed to farmers through the seed gardens. Significant variation observed for bean and fruit characteristics among cacao accessions in this study also indicated the importance of on-farm collections as a valuable reservoir of genetic diversity. Some of these traits are of commercial importance and have been used as selection criteria by farmers in the choice of parent trees for raising seedlings to make new plantings and farm expansion. From this study, the complete absence or slight anthocyanin pigmentation on the ridge of the mature fruit of more than 88% of accessions indicated that most of the cacao now grown in Nigeria was apparently derived from the Amazonian Forastero origin. The preponderance of 'Cundeamour' fruit shape (76%) with slight to strong basal constriction (88%), obtuse to attenuate apex forms (95%) and intermediate to intense rugosity (86%) showed that, possibly, the Upper Amazon Forastero (UAF) 'Parinari' population, characterized by pronounced bottleneck, conspicuous apex form and the intermediate to intensely warty fruit (Bartley 2005) and to a lesser extent, the 'Nanay' population, had the most impact on cacao On the other hand, the low percentage (less than 15%) of red pigmentation in fruits, a trait associated with some 'Criollo' populations indicated that 'Criollo' and red-podded 'Trinitario' populations have, at present, only a minimal influence on field-grown cacao in Nigeria.
