Geographic Patterns of Genetic Variation among Cacao (Theobroma cacao L.) Populations Based onChloroplast Markers

The cacao tree (Theobroma cacao L.) is native to the Amazon basin and widely cultivated in the tropics to produce seeds, the valuable raw material for the chocolate industry. Conservation of cacao genetic resources and their availability for breeding and production programs are vital for securing cacao supply. However, relatively little is still known about the phylogeographic structure of natural cacao populations. We studied the geographic distribution of cpDNA variation in different populations representing natural cacao stands, cacao farms in Ecuador, and breeding populations. We used six earlier published cacao chloroplast microsatellite markers to genotype 233 cacao samples. In total, 23 chloroplast haplotypes were identified. The highest variation of haplotypes was observed in western Amazonia including geographically restricted haplotypes. Two observed haplotypes were widespread across the Amazon basin suggesting long distance seed dispersal from west to east in Amazonia. Most cacao genetic groups identified earlier using nuclear SSRs are associated with specific chloroplast haplotypes. A single haplotype was common in selections representing cacao plantations in west Ecuador and reference Trinitario accessions. Our results can be used to determine the chloroplast diversity of accessions and in combination with phenotypic assessments can help to select geographically distinctive varieties for cacao breeding programs.


Introduction
Cacao trees (Theobroma cacao L.) are native to the Amazon basin, and their valuable seeds are the raw material for the chocolate industry [1,2]. Cultivated by 5-6 million smallholder farmers in Latin America, Africa, and Asia, cacao crops are vitally important for local economies in these continents [3,4]. As an important tropical crop, there is a constant need to develop high-yielding and disease-resistant varieties [5][6][7]. Thus, conservation and use of cacao genetic diversity are essential not only for sustainable cultivation in producing countries, but also for diverse and growing needs of the chocolate industry and consumers [1,[8][9][10][11].
Western Amazonia is considered to be a putative center of origin of cacao based on high phenotypic [1,12,13] and genetic [2,14,15] diversity. This part of the Amazon is also a region of high biodiversity [16][17][18][19][20] that in the past served as a forest refugium favoring accumulation of high tree diversity [21,22]. Western Amazonia is also considered a center

Plant Material
Fresh leaves of 233 individual cacao trees were obtained for DNA extraction from the living collection maintained by the International Cocoa Quarantine Centre, Reading University, UK. Leaf samples of 154 cacao trees were selected based on widespread geographic origin from Amazon forests in five countries ( Figure 1) (31 locations and 1-28 individuals per location, Table S1). This material corresponds to seeds or budwood collected during different expeditions in 1938-1988, aiming to obtain plant material carrying disease resistance traits and to study the species' ecology and variability [48][49][50][51][52]. Although these plant expeditions took place in Amazonian forests, some collection sites possibly experienced pre-Columbian silvicultural management like ancient cacao cultivation or translocation [26,53,54].

Plant Material
Fresh leaves of 233 individual cacao trees were obtained for DNA extraction from the living collection maintained by the International Cocoa Quarantine Centre, Reading University, UK. Leaf samples of 154 cacao trees were selected based on widespread geographic origin from Amazon forests in five countries ( Figure 1) (31 locations and 1-28 individuals per location, Table S1). This material corresponds to seeds or budwood collected during different expeditions in 1938-1988, aiming to obtain plant material carrying disease resistance traits and to study the species' ecology and variability [48][49][50][51][52]. Although these plant expeditions took place in Amazonian forests, some collection sites possibly experienced pre-Columbian silvicultural management like ancient cacao cultivation or translocation [26,53,54].
Samples of 30 individuals of cultivated cacao originated from nine different cacao farms in the coastal valley of Ecuador (1-6 individuals per location, Table S2, Figure 1). This group of germplasm is known as Refractario cacao and includes the progeny of approximately 80 different trees selected in 1937 for their resistance to witches' broom disease [48,52,55,56].  Prance (1982), and location of cacao use in Santa-Ana La Florida approximately 5300 years ago.
Additionally, leaf samples of 49 clones used for plantations and breeding programs were analyzed. They included the following genotypes: UF (United Fruit); ICS (Imperial Collection Selection); CRU (Cocoa Research Unit); TSH (Trinidad Selected); EET (Estación Experimental Tropical); and VB (Vassoura de Bruxa). The complete list of genotypes and their breeding origin in ten countries from Latin America and the Caribbean are presented in Supplementary Table S3.  Table S2, Figure 1). This group of germplasm is known as Refractario cacao and includes the progeny of approximately 80 different trees selected in 1937 for their resistance to witches' broom disease [48,52,55,56].

DNA Extraction
To extract DNA, we used about 1 cm 2 of leaf tissue and the DNeasy 96 Plant Kit (Qiagen, Hilden, Germany). DNA was diluted 1:10 before PCR amplification.

Chloroplast DNA Markers
We tested the amplification and diversity of ten universal chloroplast markers (ccmps) [57] in eight samples from distant regions. Additionally, we tested nine chloroplast microsatellite markers developed earlier for T. cacao (CaCrSSRs) based on the cacao chloroplast genome [29]. Monomorphic amplification products were obtained with all ccmp markers, but six out of the nine cacao chloroplast markers showed clear polymorphisms, which were used then to screen the 235 samples. Four of them (CaCr2, CaCr4, CaCr5, and CaCr8) represented mononucleotide, one (CaCr1)-pentanucleotide, and one-octonucleotide (CaCr9) repeats. Although it is difficult to genotype mononucleotide repeats, it is easier to do it for chloroplast SSRs than for nuclear SSRs because they are haploid and their fragment size can be easier determined due to the lack of interference with another allele such as in diploid nuclear SSRs. Different alleles were also verified by running respective samples side-by-side during the same electrophoretic run. When the allele calling was no clear, we did repetitions to ensure a correct allele scoring.
By using 1.5% agarose gel electrophoresis single bands for all PCR reactions were visualized, and the dilution ratio for the PCR product was determined. PCR products diluted at 1:10 were run on an ABI 3130xl Genetic Analyser (Applied Biosystems, Foster City, CA, USA) with GS 500 ROX used as internal size standard.

Data Analysis
Allele size analysis was performed using GeneMapper version 4.1 (Applied Biosystems, Foster City, CA, USA). Haplotypes based on all six chloroplast markers and their frequencies were determined using the Haplotype Analysis version 1.05 software [59]. This software was also used to estimate total (H T ) and within population (H S ) haplotypic diversity, genetic differentiation (F ST ) between western and eastern natural populations, and between natural populations and cultivated cacao in Ecuador. The haplotype network to visualize the relationships among haplotypes was generated using the Network 5.0.1.1 Diversity 2021, 13, 249 5 of 17 software [60]. First, a "rdf" file was created in GenAlEx 6.5 [61], and then it was reformatted and saved in Network 5.0.1.1 as an "ych" file and used by this software as input file to generate the haplotype network using the median joining method. The created output file ("out") was saved and used for visualization [60].

Identification of Haplotypes
In total, 26 alleles were identified for the six cacao chloroplast markers in 233 samples. The CaCrSSR1 marker with a pentanucleotide repeat was the most polymorphic marker with nine alleles. Six alleles were observed at CaCrSSR5, five at CaCrSSR2, three at CaCrSSR4, and two at each CaCrSSR8 and CaCrSSR9. These alleles allowed us to detect 23 haplotypes. Table 1 shows the allele compositions of these haplotypes and their frequencies. Six samples were excluded from the data analysis due to their incomplete genotyping (Tables S1-S3).

Western Amazonia, a Center of Haplotype Diversity
We observed high cpDNA genetic diversity in western Amazonia, with 19 haplotypes detected in this region in total (Figures 2 and 3). In Peruvian Amazon, 11 haplotypes were observed in three river systems, with an average of 4.6 haplotypes. The highest number of haplotypes was found along the Amazon River at Iquitos, followed by the Marañon and Nanay rivers with haplotypes H22, H16, and H19 being the most frequent, respectively. In addition, five haplotypes were observed in the south and central part of Peru along the Urubamba and Ucayali rivers ( Figure 2). Nine haplotypes were observed in Ecuadorian Amazon along the Coca and Napo rivers, with haplotypes H9 and H5 being the most frequent, respectively. Finally, two haplotypes, H22 and H12, were observed in the Colombian Amazon ( Figure 3). A high total haplotypic diversity (H T = 0.725) and a large number of haplotypes (N h = 19) were detected in western Amazonia in 28 locations ( Figure 2 and Table 2). The within population haplotypic diversity was also high (H S = 0.676) in this area.  In contrast, only two haplotypes (H16 and H17) were found in central and eastern Amazonia ( Figure 2). Both the total (HT = 0.087) and the within population (HS = 0.037) haplotypic diversity were low in eastern Amazonia (Table 2). When compared, eastern and western cacao populations were highly divergent (FST = 0.500), since they shared only two haplotypes in this study ( Figure 2).     In contrast, only two haplotypes (H16 and H17) were found in central and eastern Amazonia (Figure 2). Both the total (H T = 0.087) and the within population (H S = 0.037) haplotypic diversity were low in eastern Amazonia (Table 2). When compared, eastern and western cacao populations were highly divergent (F ST = 0.500), since they shared only two haplotypes in this study (Figure 2).

Geographically Restricted Haplotypes in Western Amazonia
In Peru, seven geographically restricted haplotypes occurred at three locations: the confluence of the Marañon and Amazon rivers (H3, H14, H15, H18, H23), Morona River (H2) and Napo River (H11). In addition, haplotype H1 was locally common in the south of Peru along the Urubamba-Ucayali River (Figure 2).
In Ecuador, the geographically restricted haplotype H13 was observed along the Napo River. Moreover, haplotypes H5 and H9 were common in different locations along the Napo basin at the border with Colombia, and haplotype H5 was common along Upano River. Haplotype 12 was observed in the Colombian Amazon at Caquetá River (Figure 3).

Correspondence between Chloroplast and Nuclear DNA Variation
There is an agreement between the chloroplast haplotypes observed in 83 individuals (Tables S1 and S3) included in Motamayor et al. [2] and eight cacao genetic groups. Each group generally has one dominant haplotype and other related haplotypes. Results may suggest that the cacao genetic groups proposed by Motamayor et al. [2] are heterogeneous in terms of the chloroplast variation (Table 3). However, to study in detail patterns of correspondence between cpDNA and nuclear variation in cacao, sampling for cpDNA analysis should ideally include all individuals analyzed by Motamayor et al. [2].
The dominant haplotypes within the genetic groups in the Peruvian Amazon were H22 in Iquitos, H19 in Nanay, H16 in Marañon, and H20 in Contamana in central Peru. In the Ecuadorian Amazon the dominant haplotype was H9 in Curaray and H10 in Nacional, and in central-eastern Amazonia H17 in Amelonado and H16 in Guiana (Table 3).

Haplotypes in Ecuadorian Plantations
Eight haplotypes were identified in 29 samples in nine cacao plantations along the coast of Ecuador, with haplotype H10 being the most frequent one (72%) from north to southwest Ecuador (Figure 3). The geographic origin of H10 is associated with three locations in four samples: Napo basin in Ecuador (N = 2), Morona (N = 1) and Marañon (N = 1) rivers in Peru (Figure 3).
Cacao farms in Ecuador also contained haplotype H2 being the second most frequent haplotype in plantations (Table 1, Figure 3). In three locations totaling three samples, H2 was observed along Napo River in Ecuador and Morona River in Peru. However, H2 and H10 may also occur in other Amazonian cacao populations that were not analyzed here.
H4 and H8 were rare haplotypes in Ecuadorian plantations observed only in Hacienda Balao and Vuelta Larga, respectively. These haplotypes were not observed in any natural population studied here (Figures 2 and 3).
Low total (HT = 0.116) and within population (Hs = 0.075) haplotypic diversity were observed in plantations of western Ecuador due to the high frequency of H10 and relatively low sample size in plantations (N = 29). Differentiation between cultivated and samples collected in Amazonian forest was high (FST = 0.500).   in Theobroma cacao L. populations. Each circle represents a specific haplotype, and its size reflects the respective haplotype frequency. Vertical bars in the lines connecting haplotypes indicate hypothetical mutations separating haplotypes. The distance between neighboring vertical lines is the same, but it was condensed in some linages for better space use and visualization of the network.   in Theobroma cacao L. populations. Each circle represents a specific haplotype, and its size reflects the respective haplotype frequency. Vertical bars in the lines connecting haplotypes indicate hypothetical mutations separating haplotypes. The distance between neighboring vertical lines is the same, but it was condensed in some linages for better space use and visualization of the network.
Low total (H T = 0.116) and within population (Hs = 0.075) haplotypic diversity were observed in plantations of western Ecuador due to the high frequency of H10 and relatively low sample size in plantations (N = 29). Differentiation between cultivated and samples collected in Amazonian forest was high (F ST = 0.500).

Haplotypes Observed in Breeding Populations and Cultivars
Cacao breeding has been mostly relying on vegetatively propagated accessions, such as Imperial College Selection (ICS), Iquitos Mixed Calabacillo (IMC), Nanay (NA), Parinari (PA), Pound, Scavina (SCA), and United Fruit (UF) [62][63][64]. Table 4 presents haplotypes detected in reference accessions and their potential geographic sources. For example, SCA 6, Pound 18, PA 139, and IMC 67 showed H5, H16, H20, and H22 haplotypes, respectively, and the geographic origin of these accessions ranged from central to northern Peru ( Figure 2). In addition, cultivars, such as UF 273 (Nacional × Amelonado), ICS, and EET (Nacional × Unknown) revealed the common haplotype H10 observed in Ecuadorian plantations (Figure 3). The cacao cultivar CCN 51 was characterized by the unique haplotype H21. This haplotype is separated by one mutational step from H22 in the mononucleotide motif repeat of CaCrSSR2 (Figure 4). H22 was also found in the International Cacao Collection in Catie, including the Matina accession, which is also a common cultivar. Another unique and closely related haplotype is H23, which was observed along with haplotype 22 in Iquitos, Peru. H22 and H23 formed a separate cluster (lineage) in the haplotype network ( Figure 4).

Distribution of Haplotype Diversity
High haplotype diversity was observed in western Amazonia (Figure 3). This region represents a hot spot of cacao diversity [10,65], where cacao populations have a high haplotype diversity in the river systems of Marañon, Amazon (Iquitos), Nanay, and Uyacali in Peru, and along the Coca and Napo river systems in Ecuador (Figures 2 and 3).
Our results support the hypothesis of decreasing cacao diversity from the western part to the eastern part of the range. The same differentiation pattern from west to east was observed in the Amazon basin for 200 cacao accessions representing ten genetic groups based on whole genome sequencing and multidimensional scaling analysis [15]. Thus, the distribution of haplotype diversity is consistent with the suggested origin of the species in western Amazonia [1,2,12,13].
Our results support the hypothesis of western Amazonia being considered as a forest refugial area [21,66] and a center of crop genetic diversity [26,67], where cacao populations could have been restricted to temporarily isolated refugia during the Pleistocene (Figure 1). After the Pleistocene, suitable habitats during the Holocene allowed migration of cacao populations by human dispersal from western to central Amazonia along the Amazon River [65]. Cacao populations may have experienced selection in this center of crop diversity to some degree by domestication for fruit pulp and later were dispersed through the basin by pre-Columbian human expansion [24,26]. Western Amazonia may also be a center of domestication which is reflected in the presence of predominant haplotypes H10, H20, and H22 both in breeding populations and plantations in western Amazonia (Figure 2 and Table 3).
Haplotypes that were observed only in western Amazonia (H12 in Colombia, H13 in Ecuador, and H11, H3, H2, H14, H15, H18, and H23 in Peru) probably are associated with cacao populations that experienced isolation in forest refugia when the Amazon forest was reduced during the Pleistocene (Figure 1) [21,22,65]. Another possible explanation assumes in situ occurrence due to edaphic adaptation to rich soils of refugial areas, which is common in tree species of Amazonia [68,69].
Although the cacao samples studied here were distributed across the Amazon basin, they mainly represented the Upper Amazon ( Figure 1). Thus, it is important to analyze additional cacao populations within the species range that could reveal new and distinctive haplotypes. For example, a recent forest survey of 1170 plots in Amazonia identified cacao as a common tree in the basin; specifically, cacao populations were observed in southwestern and southern Amazonia, and French Guyana [27,70]. Pre-Columbian human societies likely enriched these forest areas with cacao trees brought from the center of cacao diversity in western Amazonia [27].
The occurrence of only a single haplotype H16 in French Guiana could indicate human meditated dispersal of a few cacao individuals to this region. However, additional evidence to support human-mediated distribution of cacao in Amazonia is needed using more extensive sampling, specific studies such as radiocarbon and stable isotope analyses, and archaeological and ecological surveys looking at past human interactions with tropical tree species [71,72].

Chloroplast Haplotypes Match Genetic Groups Based on nSSRs
Haplotype diversity and distribution observed in our study partly mirrored the genetic clusters of the dendrogram based on 96 nSSRs (Figure 2 in [2]). For example, the Marañon, Guiana and Amelonado cluster is represented by related haplotypes 16 and 17 in the haplotype network (Table 4, Figure 4). Moreover, the cluster comprising Nanay and Iquitos is represented by the related haplotypes 19 and 22, and the closely related clusters Curaray and Nacional by the related haplotypes H10 and H9 in our haplotype network (Table 4, Figure 4).
The presence of specific haplotypes in different geographic regions and within different genetic groups suggest particular patterns of seed dispersal. For example, haplotype H20 was observed in Curaray (north of Ecuador), Iquitos (north of Peru), and Contamana (central Peru) groups ( Figure 3, Table 4). Similarly, haplotype H19 was observed in the distantly related Contamana (central Peru) and Nanay (north of Peru) groups ( Figure 3, Table 4). This may be explained by the presence of temporarily separated forest refugia in western Amazonia [21,67] where cacao populations were likely restricted during the Pleistocene [65]. Later, during the Holocene cacao populations experienced habitat expansion when the climate became wetter and warmer, favorable for wide distribution in the Amazon basin [65]. Cacao habitat expansion aided by human meditated dispersal during the Holocene times may explain ample seed dispersal observed among cacao groups [27,65].
Marañon, Guiana and Amelonado clustered together in the nSSR-based dendrogram [2]; these populations share H16, and Amelonado and Marañon also have the related haplotype H17 in common (Table 4). These haplotypes reveal a wide distribution from western to eastern Amazonia ( Figure 2) and likely experienced human dispersal in the basin. Indeed, current abundance of cacao populations in south-southwestern Amazon and French Guyana was explained partially by pre-Columbian human dispersal [27].
Cacao populations in French Guyana probably originated from local forest refugia during the Pleistocene-Holocene epochs [73]. However, an alternative explanation of their origin may be human meditated seed dispersal from western to eastern Amazon [14,27,65]. The observed wide distribution of haplotypes of H16 and H17 in the Amazon basin supports this hypothesis (Figures 2 and 3). However, to clarify the origin of Amelonado and Guiana groups more extensive sampling across the species' geographic range is needed accompanied by historical records of cacao dispersal in the Amazon. However, difficulties may arise to identify the hypothetical origin of native Amazonian crops due to forest expansion during the Holocene [24].
In the case of the Nacional group, two main natural populations with H10 were observed: one in the Ecuadorian Amazon (Napo River) and another in the Peruvian Amazon (Morona River) (Figure 2). We consider that additional cpDNA analyses in Ecuadorian populations are needed to provide more evidence about the origin of the Nacional variety. However, our results are in agreement with [2] who found 13 individuals from the Morona River clustered in the Nacional group. Furthermore, based on the nSSR genetic analysis of 65 wild individuals widespread in the Amazonian forest of Ecuador and eight individuals from the Morona River in Peru, [74] suggested an ancestral origin of the Nacional group in the Southern region of Ecuadorian Amazon close to Morona River.
The distribution of H10 suggests a pattern of seed dispersal among cacao populations in western Amazonia, which supports the hypothesis of a shared origin of this haplotype from Ecuadorian and Peruvian populations and later their introduction to the coast of Ecuador, where they were established in pure plantations. Furthermore, chloroplast analysis of Chuncho, a cacao population from Peru that resembles the Nacional fine-flavor cacao [75], would help to trace the geographic origin of fine-flavor cacao to the south of Peru as suggested by [75].

Chloroplast Haplotypes and Domesticated Cacao
Combined analyses of nuclear and cpDNA variation can help to unravel the history of cacao domestication and to narrow down the origin of domesticated varieties. For example, the Criollo variety (not included in the present study) was domesticated from a fraction of the ancestral Curaray population about 3600 years ago [15]. Within the Curaray population we observed haplotypes H9, H13, and H20 in northern Ecuador (Figure 3 and Table 4). CpDNA analyses of the Criollo variety could reveal its geographic origin within the distribution range of Curaray. Indeed, the geographic origin of haplotype H9 observed with high frequency in the Curaray group ( Figure 4) possibly suggests that trees from the Ecuador-Colombia border served as a source of plant material during the domestication of Criollo cacao (Figure 3).
Further questions arise such as why the domestication of cacao started with the Ecuadorian Curaray group and not with other populations. It was probably because its sweet pulp was used for the elaboration of fermented beverages, which may also have favored its selection and dispersal out of the Ecuadorian Amazon [24,74,76].

Haplotype Diversity in Plantations and Breeding Populations
The high haplotype diversity observed in natural populations contrasts with the low diversity observed in cacao plantations (Table 2). We observed a high contribution of haplotype H10 in both selections made in cacao plantations in the west of Ecuador and breeding populations such as the ICS clones of Trinidad (Table 3). In addition, H10 was also observed in Colombia and Grenada, providing further evidence for introduction events of this haplotype to other countries.
Cacao cultivation in coastal Ecuador has been recorded since the 17th century, and cacao genotypes named Nacional were selected and cultivated because of their special chocolate flavor [1]. Currently, new sources of genetic material are required to improve flavor, increase yield and add disease resistance traits to the Nacional cultivar [74]. We suggest that the four wild provenances associated with haplotype H10 in western Amazonia of Peru and Ecuador could be screened for agronomic traits to test their use in breeding programs, potentially adding useful traits to Nacional cultivars. The presence of six haplotypes, with haplotype H10 being the most frequent (42%), in 15 reference Trinitario accessions suggests seed introduction events occurring probably from northeast of Ecuador and north of Peru to Trinidad (Table 3). Similar to our results, five different haplotypes were observed in 21 different reference Trinitario accessions based on nine cacao cpSSRs markers [45].
The river basins associated with geographically restricted haplotypes (Figures 2 and 3) could serve as source of specific adaptations and potential new traits for cacao breeding programs, but only if these samples carry favorable agronomic traits. Examples of these areas include the Napo basin of Ecuador (H5, H9, and H13), the Caquetá River (H12) in Colombia, and Ucayali River in south and central Peru (H1).

Conclusions
Western Amazonia contains high haplotype diversity areas in Peru and Ecuador. Populations with at least two common haplotypes may have been dispersed by humans from this center of cacao genetic diversity to new suitable habitats following a west to east route of migration in the Amazon basin. The western Amazonia has high value for cacao conservation considering its haplotype diversity and presence of geographically unique haplotypes, which were not observed in breeding populations and cultivars yet and could be evaluated for agronomic traits in support of cacao improvement.
The cacao cpDNA haplotypes observed here can be used to determine the chloroplast diversity of accessions and select distinctive haplotypes in cacao breeding programs. Additionally, the map of cpDNA haplotypes can be used to verify the geographic origin of planting material at finer geographic scale and point to potential areas for new collections. Craft chocolate makers can use this reference map to create new markets for cacao products based on geographic origin.

Outlook
New collections in Amazonia will expand the availability of cacao genetic resources and help us better understand the patterns of geographic genetic variation and dispersal. However, rapid land use changes and gold mining in the Amazon forest are threatening the natural habitat of wild cacao populations and eroding its genetic diversity [8,70,77].
Supplementary Materials: The following are available online at https://www.mdpi.com/article/ 10.3390/d13060249/s1, Table S1: Clone name, ICQC (The International Cocoa Quarantine Centre) accession number, geographic origin, samples per location, genetic group according to Motamayor et al. 2008, and chloroplasts haplotypes observed in a sample of Theobroma cacao L. accessions. The haplotype column shows the order of the markers CaCr1, CaCr2, CaCr4, CaCr5, CaCr8 and CaCr9, and the fragment sizes are given in base pairs (bp). Details about donor collections, geographic location of collections expeditions and agronomic traits of these cacao clones are available on the ICGD (The International Cocoa Germplasm Database) website [52]. Table S2: Clone name, ICQC accession number, locations, and chloroplast haplotypes of Theobroma cacao L. in nine farms of the pacific coast of Ecuador. The haplotype column shows the order of the markers CaCr1, CaCr2, CaCr4, CaCr5, CaCr8 and CaCr9, and the fragment sizes are given in base pairs (bp). Collection and agronomic details of these cacao clones are available on the ICGD website. Table S3: Clones names, ICQC accession number, origin of samples, genetic group (Motamayor et al. 2008), and chloroplasts haplotypes observed in a sample of Theobroma cacao L. genotypes used for cultivation or breeding. The haplotype column shows the order of the markers CaCr1, CaCr2, CaCr4, CaCr5, CaCr8 and CaCr9, and the