Abstract
Fruit oil content (FOC) is one of the most important commercial traits in oil palm; however, extensive study on related traits is still limited. The present study was conducted to analyze the relationship between FOC and fruit-related traits, as well as to predict the oil palm germplasm for potential improvement. In this study, a total of 11 traits, including fruit bunch number (FBN), average fruit weight (AFW), mesocarp-to-fruit ratio (M/F), kernel-to-fruit ratio (K/F), shell-to-fruit ratio (S/F), average fruit length (AFL), average fruit width (AFWD), average shell thickness (AST), mesocarp oil content (MOC), kernel oil content (KOC), and FOC were analyzed in 39 germplasms collected from seven different countries in Asia and Africa. Different statistical analyses were conducted to evaluate the relationship between FOC and fruit-related traits. Correlation analysis showed that FOC was positively and significantly correlated with M/F, MOC, and KOC, whereas negatively and significantly correlated with S/F and AST. Likewise, path analysis indicated that M/F and MOC have high positive direct effect on FOC, whereas S/F and AST have high negative direct and indirect effects on FOC. Furthermore, regression analysis showed significant correlation between predicted and observed FOC. In conclusion, FOC was mainly determined by M/F, MOC, S/F, and AST, and the FOC prediction in this study was reliable for germplasm evaluation. In addition, G39 (Tenera) and G2 (Parthenocarpy) have the highest FOC with 58.62% and 57.68%, respectively, indicating that they might be potential candidates for FOC improvement. These results could be applicable to oil palm breeding programs.
African oil palm (Elaeis guineensis Jacq.) is an important woody oil crop that is widely cultivated in tropical regions all over the world (Adjei-Nsiah et al., 2012; Gutiérrez-Vélez et al., 2011; Murphy, 2009, 2014; Sayer et al., 2012). Oil palm originates from Africa, but Malaysia introduced oil palm in the 1900s and quickly developed oil palm plantations. Indonesia, Malaysia, and Thailand are currently the most important oil palm plantation regions in the world (Corley and Tinker, 2015). Current yields have generally reached ≈4 to 6 tons per hectare, which led to a rapid expansion in the tropics of Asia, Africa, and South America. Supply of palm oil and palm kernel oil accounted for 39.7% of vegetable oil around the world (U.S. Department of Agriculture, 2018). Palm oil and palm kernel oil are crucial raw materials for the food industry, chemical industry, and biodiesel (Edem, 2002; Kalam and Masjuki, 2002; Mba et al., 2015).
Although oil palm was introduced to China in 1926, it has not been planted commercially because of low oil yield in China (Shi et al., 2015). Hence, China imports more than 5 million tons of palm oil every year (Xiong et al., 2009). However, such import can be reduced by developing high oil-yielding germplasms and breeding high-yield cultivars. The evaluation of factors most related to oil yield can be useful in selecting and developing germplasms with high oil yield.
Generally, oil palm fruit can be divided into exocarp, mesocarp, endocarp (shell), and kernel (Corley and Tinker, 2015). The fruit components, including fruit weight, fruit size, M/F, K/F, and S/F, might be crucial for FOC in oil palm germplasm. The AFW of high-yielding Thai and Nigerian cultivars is ≈7.57 g (Owolarafe et al., 2007; Patcharin et al., 2013; Rajanaidu et al., 1979); however, high variations were observed in M/F, S/F, and K/F (Krualee et al., 2013; Patcharin et al., 2013; Rajanaidu et al., 1979; Sapey et al., 2012). Mesocarp and kernel of the fruit are the richest sources of oil, and it is relatively important to increase the oil content of mesocarp and kernel of existing germplasms in China. To our knowledge, previous studies (Krualee et al., 2013; Mathews et al., 2004; Okoye et al., 2009; Rafii et al., 2013) mainly focused on oil yield from the whole bunch of the oil palm fruits. Although the fruit components of oil palm have been extensively studied, investigations on FOC are still limited. Breeders often use different statistical analyses to estimate the interrelation among the agronomic traits (Boo et al., 1990). Hierarchical cluster analysis shows the similarity of the samples (Peeters and Martinelli, 1989). Path analysis serves as a great tool to evaluate the relationship among agronomic traits (Krualee et al., 2013; Patcharin et al., 2013; Zeng et al., 2016).
In the present study, a total of 39 germplasms, including Dura, Tenera, and Parthenocarpy, collected from seven different countries in Asia and Africa were used, and 11 fruit-related traits among these germplasms were observed and then analyzed using cluster, analysis of variance (ANOVA), correlation, path, and regression analysis, respectively. Furthermore, FOC prediction was constructed and validated for germplasm evaluation. The expected results could be applicable to oil palm breeding programs.
Materials and Methods
Plant materials.
A total of 39 oil palm germplasm materials (9 years old), representing seven countries in Africa and Southeast Asia (Table 1), maintained at Coconut Research Institute, Chinese Academy of Tropical Agricultural Sciences, Wenchang, China, were used in this study. The collected germplasms included three fruit forms: Dura, Tenera, and Parthenocarpy (Fig. 1). Each germplasm has only one tree; fully matured fruits (fall off easily from bunch according to the oil palm harvesting standards) from each germplasm were collected for the following analyses.
Origin and fruit forms of collected materials.
The fruit forms of collected oil palm. (A) Dura (thick shell). (B) Tenera (thin shell). (C) Parthenocarpy (without embryo and endosperm).
Citation: HortScience horts 54, 8; 10.21273/HORTSCI14044-19
The fruit forms of collected oil palm. (A) Dura (thick shell). (B) Tenera (thin shell). (C) Parthenocarpy (without embryo and endosperm).
Citation: HortScience horts 54, 8; 10.21273/HORTSCI14044-19
The fruit forms of collected oil palm. (A) Dura (thick shell). (B) Tenera (thin shell). (C) Parthenocarpy (without embryo and endosperm).
Citation: HortScience horts 54, 8; 10.21273/HORTSCI14044-19
Studied traits.
Statistical analyses.
Phenotypic coefficient of variation (PCV), one-way ANOVA, correlation coefficients, path analysis, and regression analysis were calculated using SAS Statistical Package Program version 9.1 (SAS Institute, Inc., Cary, NC) (Each, 2006; Ziegel, 1996). The diagrams for cluster analysis, box plot, and ANOVA were generated by OriginPro 9.0.
Results
Cluster analysis.
According to cluster analysis, three main clusters can be separated. The germplasm G2 and G13, which were Parthenocarpy from Thailand and Malaysia, were in one cluster, and G11, G29, G19, G12, and G31, which included Dura and Tenera from Malaysia and Nigeria, were grouped in one cluster. The remaining 32 germplasms, which consisted of 31 Tenera from seven different countries and one Dura from Cameroon, formed the largest cluster. Clustering analysis showed that Tenera and Dura were more closely related than to Parthenocarpy, and the overall results indicated the close association of germplasm from different countries (Fig. 2).
Cluster dendrogram of 39 germplasms.
Citation: HortScience horts 54, 8; 10.21273/HORTSCI14044-19
Cluster dendrogram of 39 germplasms.
Citation: HortScience horts 54, 8; 10.21273/HORTSCI14044-19
Cluster dendrogram of 39 germplasms.
Citation: HortScience horts 54, 8; 10.21273/HORTSCI14044-19
Phenotypic variation of analyzed traits.
The PCVs ranged from 10.46% to 78.28%. The highest PCV was recorded for AST (78.28%) followed by S/F (57.12%), AFW (35.59%), K/F (35.43%), FBN (28.92%), FOC (19.44%), and KOC (16.05%), whereas the lowest PCV was observed in AFL (10.46%) (Table 2; Fig. 3). Most of the traits varied largely, indicating that germplasms collected from seven countries had abundant diversity.
Phenotypic variation of analyzed traits among 39 germplasms.
Box plot showing the variation of observed traits. FBN = fruit bunch number; AFW = average fruit weight; M/F = mesocarp-to-fruit ratio; K/F = kernel-to-fruit ratio; S/F = shell-to-fruit ratio; AFL = average fruit length; AFWD = average fruit width; AST = average shell thickness; MOC = mesocarp oil content; KOC = kernel oil content; FOC = fruit oil content.
Citation: HortScience horts 54, 8; 10.21273/HORTSCI14044-19
Box plot showing the variation of observed traits. FBN = fruit bunch number; AFW = average fruit weight; M/F = mesocarp-to-fruit ratio; K/F = kernel-to-fruit ratio; S/F = shell-to-fruit ratio; AFL = average fruit length; AFWD = average fruit width; AST = average shell thickness; MOC = mesocarp oil content; KOC = kernel oil content; FOC = fruit oil content.
Citation: HortScience horts 54, 8; 10.21273/HORTSCI14044-19
Box plot showing the variation of observed traits. FBN = fruit bunch number; AFW = average fruit weight; M/F = mesocarp-to-fruit ratio; K/F = kernel-to-fruit ratio; S/F = shell-to-fruit ratio; AFL = average fruit length; AFWD = average fruit width; AST = average shell thickness; MOC = mesocarp oil content; KOC = kernel oil content; FOC = fruit oil content.
Citation: HortScience horts 54, 8; 10.21273/HORTSCI14044-19
With respect to FOC, ANOVA showed that FOC from 39 germplasms was significantly different (P < 0.0001). The lowest FOC was observed in G12 (20.44%), whereas the highest values were represented by G39 (58.62%), which was Tenera from Angola, and G2 (57.68%), which was Parthenocarpy from Thailand (Fig. 4). Germplasms G39 and G2 had the potential of breeding high oil-yielding varieties.
Analysis of variance for fruit oil content (FOC) in 39 germplasms.
Citation: HortScience horts 54, 8; 10.21273/HORTSCI14044-19
Analysis of variance for fruit oil content (FOC) in 39 germplasms.
Citation: HortScience horts 54, 8; 10.21273/HORTSCI14044-19
Analysis of variance for fruit oil content (FOC) in 39 germplasms.
Citation: HortScience horts 54, 8; 10.21273/HORTSCI14044-19
Correlation among analyzed traits.
Correlation analysis showed that FOC was positively and significantly correlated with MOC (r = 0.74**), M/F (r = 0.66**) and KOC (r = 0.34*), whereas negative and significant correlations of FOC were observed with S/F (r = −0.69**) and AST (r = −0.61**). Significant negative correlations of M/F with K/F (r = −0.50**), S/F (r = −0.96**), AFWD (r = −0.35*), and AST (r = −0.69**) was also found in this study. Positive and significant correlation existed between AFW and AFL (r = 0.76**). In addition, S/F was positively and significantly correlated with AFWD (r = 0.33*) and AST (r = 0.81**) (Table 3). Accordingly, FOC may be related mostly with MOC, M/F, KOC, S/F, and AST.
Correlation coefficients of analyzed traits.
Path analysis.
The direct and indirect effects of various traits on FOC were estimated using path analysis. Results showed that M/F (0.813) had the highest positive direct effect on FOC followed by MOC (0.731). The positive direct effect of KOC (0.073) on FOC was little and ignorable. Although the direct effect of S/F (0.266) on FOC was not obvious, S/F had considerable negative indirect effect on FOC through M/F (−0.780). Similarly, AST had negative direct (−0.102) and indirect effects on FOC through M/F (−0.560). The effects of other traits on FOC were ignorable because of the little amount and also the correlation was not significant (Table 4), indicating that FOC was mainly determined by M/F, MOC, S/F, and AST.
The direct and indirect effects of fruit-related traits on FOC.
Regression analysis between FOC and fruit-related traits.
Results showed that the correlation coefficient between predicted and observed FOC was significant (r = 0.99) (Table 5). It indicated that it is reliable to predict FOC by using the regression equation. Among all these germplasms, G39 was found to have the highest FOC and followed by G2, which was in agreement with the observations (Fig. 4).
Prediction values and observation values of collected germplasm.
Discussion
Oil palm is the highest oil-yielding crop among oleaginous crops and producing 4 to 6 tons of oil per hectare on average. With the increasing population and oil consumption, it is estimated that the demands for edible oil will hit 240 million tons by 2050 (Barcelos et al., 2015). To overcome such exigencies for edible oil, it is necessary to improve the FOC of the potential oil palm cultivars through effective breeding programs. There is an enormous potential for increasing oil yield of more than 10 tons per hectare by selecting traits most related to oil yield in breeding programs. Evaluation and selection of potential oil palm individuals from the existing germplasm is important to achieve high oil yields, such as 11 to 18 tons per hectare. Breeders often use different statistical methods to determine the relationship among the agronomic traits (Oboh and Fakorede, 1990).
In the current study, interrelation among different traits as well as direct and indirect effects of specific traits on FOC was estimated using different statistical analyses. Hierarchical cluster analysis shows the relationship among the samples (Ene et al., 2016). Clustering analysis in this study revealed that Tenera and Dura were more closely related than to Parthenocarpy. Tenera is popular as a result of crossing between Dura × Pisifera, and hence Tenera and Dura had higher similarity as compared with Parthenocarpy. Cluster analysis also indicated the close association of germplasms from different countries. Oil palm populations from Africa were introduced to many places by humans (Hayati et al., 2004), and it is most likely that there is no dramatic variation during its spreading to Asia from Africa.
The cv determines the relative amount of variability among the traits (Sharma, 1988). The amount of variability present determines the potential in trait selection. The traits with greater variability have the higher potential to discriminate among treatments (Khan et al., 2009). The results of the PCV in the current study showed that the highest PCV was recorded for AST followed by S/F, AFW, K/F, FBN, FOC, and KOC, indicating the presence of high variability and hence there is a wider scope of selection in these traits as compared with other traits. The ANOVA with respect to FOC of 39 germplasms showed that G39 (Tenera) and G2 (Parthenocarpy) had the highest FOC, suggesting that G39 and G2 might be potential candidates for FOC improvement.
Both mesocarp and kernel in oil palm are the rich sources of oil, and hence the study on mesocarp and kernel as well as the factors affecting the accumulation of the oil in these parts is relatively important. Previous studies showed that M/F in Thailand ranged from 47.90% to 73.81% (Krualee et al., 2013); however, in the present study, M/F in Hainan Island was 74.26%, although the climate conditions such as temperature and rainfall in Hainan Island were inferior to southern Thailand. Similarly, the fruit characters, including M/F, AFL, AFWD, and AST in Hainan Island were better when compared with those in Nigeria and Ghana (Owolarafe et al., 2007; Rajanaidu et al., 1979; Sapey et al., 2012).
Interrelations of various traits in oil palm were discussed in previous studies (Oboh and Fakorede, 1990; Rios et al., 2018); however, extensive study on FOC, which is one of the major components responsible for oil yield, is still limited. Correlation analysis in this study showed a significant negative correlation of M/F with S/F, K/F, and AST. Similar reports have been presented in previous studies (Okoye et al., 2009; Patcharin et al., 2013). This negative correlation suggests that M/F is negatively influenced by S/F, K/F and AST. The increase in kernel and shell size would result in a decrease in mesocarp content (Okoye et al., 2009). In addition, FOC was positively and significantly correlated with M/F, MOC, and KOC, and negatively and significantly correlated with S/F and AST. These results suggested that increasing the M/F, MOC, and KOC will lead to higher FOC. In contrast, reducing the S/F and AST can increase the possibility of improving FOC.
Path analysis was used to determine the direct and indirect effects of various fruit components on FOC (Wright, 1921). Previous studies (Krualee et al., 2013) reported that M/F has a direct effect on oil yield per palm and also has an effect on the percentage of oil per bunch. Path analysis in the current study revealed that M/F and MOC have high positive direct effects on FOC. The direct effect of KOC on FOC was positive but with little amount. The indirect negative effect of S/F on FOC through M/F is also relatively high. Furthermore, the high negative direct and indirect effects of AST on FOC were also observed in this study. The results of path analysis were in agreement with that of correlation analysis. Regression analysis also confirmed that FOC was positively related to M/F and MOC, and negatively related to S/F and AST. The significant correlation (r = 0.99) between predicted and observed FOC in regression analysis suggested that this regression equation is reliable for evaluation on FOC in oil palm germplasms. These observations revealed that M/F, MOC, S/F, and AST strongly influence FOC and hence should be preferably used as selection criteria for development of high-yield oil palm cultivars.
In conclusion, FOC was mainly determined by M/F, MOC, S/F, and AST, and the FOC prediction in this study was reliable for germplasm evaluation, which could be applicable in oil palm breeding programs.
Literature Cited
Adjei-Nsiah, S., Sakyi-Dawson, O. & Kuyper, T.W. 2012 Exploring opportunities for enhancing innovation in agriculture: The case of oil palm production in Ghana J. Agr. Sci. 4 10 1275 1279
Barcelos, E., Rios, S.D.A., Cunha, R.N.V., Lopes, R., Motoike, S.Y., Babiychuk, E., Skirycz, A. & Kushnir, S. 2015 Oil palm natural diversity and the potential for yield improvement Front. Plant Sci. 6 190
Boo, M.H., Park, C.Y. & Jin, W.K. 2005 Finite element analysis of deformation and restoration for cylindrical beam structure. International Conference on Structural Mechanics in Reactor Technology
Corley, R. & Tinker, P. 2015 The oil palm. John Wiley & Sons, Chichester, UK
Each, O. 2006 SAS 9.1.3 language reference: dictionary [M]. SAS Institute, Cary, NC
Edem, D. 2002 Palm oil: Biochemical, physiological, nutritional, hematological and toxicological aspects: A review Plant Foods Hum. Nutr. 57 3-4 1275 1279
Ene, C.O., Ogbonna, P.E., Agbo, C.U. & Chukwudi, U.P. 2016 Studies of phenotypic and genotypic variation in sixteen cucumber genotypes. Chil. J. Agr. Res. 76(3):307–313
Gutiérrez-Vélez, V.H., DeFries, R., Pinedo-Vásquez, M., Uriarte, M., Padoch, C., Baethgen, W., Fernandes, K. & Lim, Y. 2011 High-yield oil palm expansion spares land at the expense of forests in the Peruvian Amazon Environ. Res. Lett. 6 4 044029
Hayati, A., Wickneswari, R., Maizura, I. & Rajanaidu, N. 2004 Genetic diversity of oil palm (Elaeis guineensis Jacq.) germplasm collections from Africa: Implications for improvement and conservation of genetic resources Theor. Appl. Genet. 108 1274 1284
Kalam, M. & Masjuki, H. 2002 Biodiesel from palm oil—an analysis of its properties and potential Biomass Bioenergy 23 6 1275 1279
Khan, A.S.M.M.R., Kabir, M.Y. & Alam, M.M. 2009 Variability, correlation, path analysis of yield and yield components of pointed gourd. J. Agr. Rural Dev. 7(1-2):93–98
Krualee, S., Sdoodee, S., Eksomtramage, T. & Sereeprasert, V. 2013 Correlation and path analysis of palm oil yield components in oil palm (Elaeis guineensis Jacq.) Witthayasan Kasetsat Witthayasat 47 528 533
Mathews, J., Lee, A.K., Clarence, P.J., Chung, M.Y. & Rao, S. 2004 Oil content in oil palm fruit mesocarp and bunch, and some of its related physiological and agronomical factors Planter 80 938 1275 1279
Mba, O.I., Dumont, M.J. & Ngadi, M. 2015 Palm oil: Processing, characterization and utilization in the food industry–A review Food Biosci. 10 26 41
Murphy, D.J. 2009 Oil palm: Future prospects for yield and quality improvements Lipid Technol. 21 11-12 1275 1279
Murphy, D.J. 2014 The future of oil palm as a major global crop: Opportunities and challenges J. Oil Palm Res. 26 1 1275 1279
Oboh, B.O. & Fakorede, M.A.B. 1990 Interrelations among vegetative, yield and bunch quality traits in short-stem oil palm progenies Euphytica 46 7 14
Okoye, M., Okwuagwu, C. & Uguru, M. 2009 Population improvement for fresh fruit bunch yield and yield components in oil palm (Elaeis guineensis Jacq.) Amer.-Eurasian J. Sci. Res. 4 59 63
Owolarafe, O., Olabige, M. & Faborode, M. 2007 Physical and mechanical properties of two varieties of fresh oil palm fruit J. Food Eng. 78 4 1275 1279
Patcharin, T., Yaowanat, H., Puntaree, T. & Srinives, P. 2013 Estimates of repeatability and path coefficient of bunch and fruit traits in Bang Boet dura oil palm J. Oil Palm Res. 25 1 1275 1279
Peeters, J.P. & Martinelli, J.A. 1989 Hierarchical cluster analysis as a tool to manage variation in germplasm collections Theor. Appl. Genet. 78 1 1275 1279
Rafii, M., Isa, Z., Kushairi, A., Saleh, G. & Latif, M. 2013 Variation in yield components and vegetative traits in Malaysian oil palm (Elaeis guineensis jacq.) dura × pisifera hybrids under various planting densities Ind. Crops Prod. 46 147 157
Rajanaidu, N., Arasu, N.T. & Obasola, C.O. 1979 Collection of oil palm (Elaeis guineensis Jacq.) genetic material in Nigeria, 2: Phenotypic variation of natural population MARDI Res. Bull. 7 1 27
Rios, S.D.A., Cunha, R.N.V.D., Lopes, R., Barcelos, E., Rocha, R.N.C.D. & Lima, W.A.A.D. 2018 Correlation and path analysis for yield components in Dura oil palm germplasm Ind. Crops Prod. 112 724 733
Sapey, E., Adusei-Fosu, K., Agyei-Dwarko, D. & Okyere-Boateng, G. 2012 Collection of oil palm (Elaeis guineensis Jacq.) germplasm in the northern regions of Ghana Asian J. Agr. Sci. 4 5 1275 1279
Sayer, J., Ghazoul, J., Nelson, P. & Boedhihartono, A.K. 2012 Oil palm expansion transforms tropical landscapes and livelihoods Global Food Security. 1 2 1275 1279
Sharma, J.R. 1988 Statistical and biometrical techniques in plant breeding. New Age International Limited Publishers, New Delhi, India
Shi, P., Feng, M.L., Lei, X.T. & Cao, H.X. 2015 Path and principal component analysis of yield traits in oil palm fruit bunch Guangdong Agr. Sci. 42 18 25
U.S. Department of Agriculture 2018 Major vegetable oils: World supply and distribution (commodity view). Production, supply and distribution online report. 31 Dec. 2018. <https://apps.fas.usda.gov/psdonline_legacy/>.
Wright, S. 1921 Correlation and causation J. Agr. Res. 20 7 1275 1279
Xiong, H.B., Li, R., Li, X.J., Fan, H.K. & Ma, Z.L. 2009 Investigation, analysis and the advice of palm industry in China [J] Chinese Agr. Sci. Bul. 24 114 117
Zeng, X.H., Pan, D.L., Liu, Z., Chen, J.M. & Lin, W.F. 2016 Evaluation of cold tolerant high yielding oil palm germplasm in Guangdong province of south China, a northern tropical region J. Oil Palm Res. 28 3 1275 1279
Ziegel, E. 1996 Applied multivariate statistics with SAS® Software Technometrics 38 324
