All solvents used were of analytical or HPLC grade and supplied by Merck (Darmstadt, Germany). Water was purifed in a Milli-Q water purification system (Millipore, Bedford, MA, USA). Standard compounds such as α, γ, δ-tocopherol (≥99%) and γ-tocotrienol (≥97%), were purchased from Sigma-Aldrich (St. Louis, MO, USA). The FA methyl ester (FAME) standards were obtained from Supelco (Bellefonte, USA).

Ten cocoa hybrids were obtained by hand pollination from six parental clones: two local Trinitario (SNK13 and SNK413), one Trinitario introduced from Trinidad (ICS40), and three Forestero (SCA12, UPA134 and T79/501) from the gene banks of the Cameroon Cocoa Development Corporation (SODECAO) at the Mengang Station (situated at 82 km from Yaounde; 3˚53' NL, 12˚03' EL in the centre region of Cameroon). Mengang station falls under a bimodal forestry zone recording an annual rainfall of 1500 - 2000 mm and an average temperature of 25˚C. These clones were selected for their tolerance to Phytophthora megakarya and their productivity [27] [28] . Crossings were conducted over two cocoa crop years (from February 2 to November 21, 2017 and from February 9 to November 29, 2018) as describe by [25] .

One thousand ripe cocoa pods from different hybrids were harvested from the experimental plots of the SODECAO at Mengang Station. After two days of pod storage, the ripe pods were split and beans obtained were fermented using the box method. The fermentation was done by putting the extracted cocoa beans in the boxes. The beans were covered with banana leaves and fermented for six days with opening and overturning every two days. The fermented cocoa beans were then lyophilized.

Total fat content was evaluated by continuous extraction by the Soxhlet method using hexane as solvent [29] . The extract was filtered and the hexane removed by evaporation at 50˚C under vacuum. Then, the obtained fats were poured into dark glass bottles, hermetically sealed and stored at −18˚C to avoid glycerides and minor compounds oxidation and/or degradation, until analyses. The result was reported in grams of fat per 100 g of dry weight (g.100g⁻¹ DW).

Fatty acid methyl esters (FAME) were prepared using the standard method according to ISO E. 5509 (2000) with modifications. 0.7 mL 10 N KOH, 5.3 mL methanol and 100 mg initial lipid extract were stirred continuously at 50˚C for one hour. 500 μL of toluene was added together with BF₃/CH₃OH, to allow good dissolvement of lipid extracts. The samples were cooled down at room temperature and 0.58 mL 24 N H₂SO₄ was added. All samples were stirred continuously at 50˚C for one additional hour. FAME were extracted with 2 mL cyclohexane, mixed and centrifuged at 9000 g for 5 minutes at 10˚C. Molecular sieves were added for 10 minutes in the final FAME solution to further neutralize the mixture. Then solution was diluted and nitrogen was blown on every sample to avoid oxidation and then the samples were directly analysed by gas chromatography.

FAME were determined according to the gas chromatography method described by [30] by capillary gas chromatography on Agilent Technologies 6890 (USA) GC device equipped with split/splitless injector, flame ionization detector (FID) and capillary column Supelco SP-2560 (length 100 m, i.d. 0.25 mm, film thickness 0.20 µm, Supelco, Bellefonte, USA). Injector and detector temperatures were 250˚C and 260˚C, respectively. Helium was used as the carrier gas at a flow rate of 5 mL∙min⁻¹. The injected volume was 1 µL and the injector split ratio was set at 20:1. The column temperature was programmed from the initial 50˚C (held 5 min) to 240˚C (held 20 min), with a temperature rate of 4˚C∙min⁻¹. The chromatographic peaks in the samples were identified by comparing the relative retention times of FAME peaks with peaks in Supelco 37 Component FAME mix standard (Supelco, Bellefonte, USA). Indeed, the content of methyl esters of fatty acids (MEMK) is determined, and the relative content of individual fatty acids (FA) is determined (calculated) accordingly.

Pure tocopherol (used as standards of tocopherol and tocotrienol) was diluted in hexane (HPLC grade), at a concentration of a 10 μg∙mL⁻¹. The analysis was performed using an UHPLC binary system (Dionex Ultimate 3000 RSLC, Waltham, MA, USA) equipped with a binary pump (HPG 3200RS), auto-sampler (WPS 3000), column compartment (TCC 3000) and the Chromeleon software for data acquisition and processing. A Nova-Pack C18 analytical column (150 × 4.6 mm i.d.) with a particle size of 2.2 μm (INTERCHIM, Uptisphere strategy SI) was used. Briefly, 80 mg of cocoa butter were dissolved in 1 mL of hexane/isopropanol (3:2 v/v), vortexed, filtered through 0.45 μm pores filters and then injected into the HPLC system. The mobile phase consists of a mixture of acetonitrile and methanol (1:1, v/v) at a constant flow rate of 1 mL∙min⁻¹ in isocratic elution. The injection volume was 20 μL. The FLD (Dionex 3400RS, Waltman, MA, USA) spectrofluorometer was used as detector, set at excitation wavelength of 295 nm and emission at 325 nm, using external standard calibration curves of α, γ, δ-tocopherol and γ-tocotrienol. Total tocopherol content was calculated as the sum of the identified and quantified tocopherols and expressed as µg∙g⁻¹ of lipid extract.

The values are means of three replications. Results were analyzed with variance (ANOVA) followed by Tukey HSD at P < 5% to compare means with the assistance of SPSS 20.0 for windows. Principal component analysis (PCA) was performed to establish associations among tocopherols and fatty acids by using SPAD 5.5 software package. Excel software was used to draw some figures.

The average TF content of the different fermented cocoa hybrids varied between 44.54% ± 1.37% and 59.32% ± 1.83% (Table 1). This highest value was recorded by T79/501 × SNK13 (Forastero female-clone). All the hybrids, except SCA12 × ICS40 and SCA12 × UPA134, were distinguished through an average fat content above 50.00%. [31] reports that in Criollo cocoa, the fat content tends to be lower than Forastero, estimated to be approximately 53%. Nonetheless, SCA12 × ICS40 and SCA12 × UPA134 (both Forastero mother-clone) recorded an average fat content of 44%. In addition, results obtained for different hybrids were close to the ranges reported by several authors for fine cocoa samples [4] [32] [15] [33] . The variations in the fat content of cocoa beans could be related to several factors such as environmental conditions, growing area [15] [31] and genotype. For this purpose, in this study, significant differences (P < 0.05) in terms of TF content were observed amongst hybrids with the same mother-clone, suggesting an impact of pollen origin. [25] assumed that, allopollen grains induce an increase in beans weight and the female-clone organizes itself to create an adequate environment for the expansion of seeds. We can suggest that this expansion is caused in part by increasing the fat content. On the other hand, [34] highlighted that the cacaos with highest market value, are those with fat contain above 56%. In this sense, only SNK413 × SCA12 (Trinitario female-clone) and T79/501 × SNK13 can be considered as having high commercial value.

The fatty acid (FA) composition in fermented samples is illustrated in Figure 1 and Table 2. The major FA according to their average content are stearic (34.57%), palmitic (26.13%), oleic (34.13%) and linoleic (3.16%) acids. The saturated fatty acids (SFA) content ranged between 59.25% ± 0.78% (for SCA12 × ICS40) and 63.7% ± 0.21% (for SCA12 × UPA134) while polyunsaturated fatty

Values with the same letter are not significantly different (P < 0.05, Tukey HSD)

Values with the same letter in the same line are not significantly different (P < 0.05, Tukey HSD).

acids (PUFA) varied from 2.8% ± 0.15% (for T79/501 × SNK413) to 4.35% ± 0.10% (for SCA12 × ICS40). SNK13 × T79/501 (hybrid with Trinitario female-clone) and SCA12 × SNK13 recorded the highest content of monounsaturated fatty acid (MUFA). [35] also noted that the chemical fatty acid composition of cocoa butter is predominantly made up of saturated fatty acids, followed by the monounsaturated ones. In general, the fatty acid composition obtained is in agreement with the results reported by [15] . These authors found average contents of palmitic (27.6%), stearic (33.8%) and oleic (34.7%) acids in Ecuadorian cocoa (known as fine cocoa) samples. The similarity between these results can suggest that our hybrids have characteristic features of fine cocoas, especially as [36] reported the similar composition of MUFA (34.06%) in dark-chocolate samples with fine Criollo cocoa from Mexico. The major MUFA in all the hybrids is oleic acid (32.3% ± 0.12 to 35.3% ± 0.20%). These one play an important role in the prevention and treatment of cardivascular diseases through different mechanisms [37] . In this sense, SNK13 × T79/501 and SCA12 × SNK13 having the healthiest FA profile. Furthermore, significant differences in terms of linoleic acid content were observed amongst hybrids with the same mother-clone (P < 0.05). The same observation was done for stearic acid content except between T79/501 × SNK13 and T79/501 × SNK413. Further investigations are needed to elucidate the role of pollen grain on the profile of these two particular fatty acids in hand-pollinated cocoa hybrids. On the other hand, others fatty acids (palmitoleic, erucic, gadoleic and α-linoleic acids) have been identified in trace amounts (<1 %) in our samples. This result is in agreement with those of [20] . According to her findings, the quantity of palmitoleic acid, linoleic acid and unusual fatty acids could be considered as markers of the fermentation status. The very low erucic acid content is an asset for the nutritional quality of cocoa butter because erucic acid is thought to be involved in harmful effects on the cardiovascular system. Although fermentation has shown to have a significant impact on the lipid profile of cocoa [20] , differences in the composition of the fatty acids profile among the fermented samples can be mainly explained by the effect of the cocoa genotype which is linked to pollen genetic origin. In addition, the UFA/SFA ratio varied from 0.57 (SCA12 × UPA134) to 0.65 (SCA12 × ICS40). This ratio determines the degree of hardness [38] and the melting properties [36] of cocoa butter. The lower this ratio, the harder the butter and the lower the melting point. This is explained by the fact that TAG with saturated fatty acids forms a more compacted structure, which is more difficult to melt. According to this point of view, SCA12 × ICS40, SCA12 × SNK13, SNK13 × T79/501 and SNK413 × SCA12 have the least hard cocoa butters.

The concentration of total tocopherols (Figure 2) ranges from 96.34 to 306.75 µg∙g⁻¹ of lipid extract of fermented cocoa samples. The results of tocopherol and tocotrienol analysis Table 3 showed a predominance of γ-tocopherols (94.64 ±

Values with the same letter in the same column are not significantly different (P < 0.05, Tukey HSD) nd: not detected.

1.51 to 292.16 ± 3.17 µg∙g⁻¹), whereas only a small amount of β and δ-tocopherol (from 0.46 to 2.78 µg∙g⁻¹ and 0.12 to 5.82 µg∙g⁻¹ respectively) was observed. An absence of α-tocopherols is also observed in three hybrids with Trinitario mother-clone (SNK13 × ICS40, SNK413 × T79/501 and ICS40 × SCA12) and three hybrids with Forastero mother-clone (SCA12 × UPA134, T79/501 × SNK413 and T79/501 × SNK13) unlike the others hybrids. The β-tocopherol content, although being minor, was comparable with that of some common edible oils [38] . The amounts of δ-tocopherol obtained in this study were lower than those reported by [39] in Quercus oils. On contrary, the β-tocopherol was found in higher amount followed by γ-tocopherol in another study conducted by [40] . It is well known that tocopherols are natural antioxidants and their presence in oilseeds is often correlated with a relative abundance of UFA [41] . Indeed, tocopherols play a role in protecting MUFA and PUFA from oxidation [38] . SCA12 × ICS40 showed the highest tocopherol content whereas his reciprocal hybrid ICS40 × SCA12, had the lowest content of both the γ-tocopherols and the total tocopherol content. This result proves a qualitative improvement of some hybrids which could be attributed to genetic mixing. Futhermore, significant differences (P < 0.05) in terms of tocopherol content were observed amongst hybrids with the same mother-clone. One of the main objectives of nutraceutical industries is to identify plant matrices rich in antioxidant compounds, for the preparation of new phytoderived products. Some hybrids studied like SNK13 × T79/501, SNK413 × SCA12, SCA12 × ICS40 and SCA12 × SNK13 seem to be the good candidates for these industries. Nevertheless, chromatographic analysis does not reveal the presence of γ-tocotrienols (Figure 3). An exploration of other forms of tocotrienols in cocoa beans should be interesting.

A PCA was performed using the content of tocopherols obtained after chromatographic analysis. A biplot of the first two principal components (PC1 = 60.12% and PC2 = 37.37%) is shown in Figure 4 for fermented beans. SNK13 × T79/501 was located at the negative axis of F1 and was correlated positively to β and δ-tocopherol while SNK413 × SCA12 and SNK413 × T79/501 were located at the positive axis of F2 and correlated to γ-tocopherol. A PCA biplot of fatty acids (PC1 = 54.61% and PC2 = 25.59%) is shown in Figure 5. Here, the clustering depicted the predominance of fatty acids (palmitic, oleic, palmitoleic, linoleic) in

three hybrids namely SNK13 × T79/501, SCA12 × SNK13 and SNK413 × SCA12, located at the postive axis of F2. Contrarily, the hybrids T79/501 × SNK13, SCA12 × UPA134 and T79/501 × SNK413 were located at the negative axis of F1. These results indicate that the fatty acid and tocopherol profile can be used to discriminate some hybrids. However, the impact of fermentation on this profile should not be neglected. The genetic origin of each hybrid could also be a major factor in the biosynthesis of these compounds. In fact, several authors have found that genotypes and post-harvest practices may affect the chemical composition of cocoa beans [15] [19] [42] .