Genetic Diversity of the Seed Physical Dimensions in Cultivated and Wild Relatives of Sesame (Genera Sesamum and Ceratotheca)

MATERIALS AND METHODS

Twenty seeds were randomly selected from the harvested seeds of each of the eight tagged plants of nine sesame accessions studied in potted field experiment in 2005 and 2006 (Table 1). The seeds of the nine accessions are as shown in Fig. 1. This investigation was carried out at the Department of Pure and Applied Biology, Ladoke Akintola University of Technology, Ogbomoso, Nigeria. The three linear dimensions of each seed namely, length (SLT), width (SWD) and thickness (STH) were measured using a micrometer screw gauge with a reading accuracy within 0.01 mm.

Table 1:	Name of sesame accessions studied, place of collection and germplasm type

Fig. 1:	Appearances of the seeds of the nine sesame accessions employed in the study

From these three linear dimensions, SLT-SWD ratio, SLT-STH ratio and SWD-STH ratio were determined. The 1000-Seed weight was measured while seed oil content (%) was determined as described by Azeez and Morakinyo (2011).

The geometric mean diameter of seed (De) which is the overall average of the three measured dimensions (i.e., Length, Width and Thickness) was determined using the equation given by Mohsenin, 1986). The seed sphericity ‘Φ’ (i.e., the extent to which the seed is spherical) was determined by the relation Φ = WT/L² while the surface area ‘S’ was determined by the relationship given by McCabe et al. (2005).

Statistical analyses: Data collected were subjected to analysis of variance, Duncan’s multiple range tests and correlation analysis on the SPSS software version 10. Principal component analysis was conducted on Statistical Analysis System Software (SAS, 2000) and a cluster analysis was performed based on unweighted pair group method using arithmetic averages (UPMGA) of Sokal and Michener (1958). The significant levels of the statistical analyses were described at the probability of 0.05 and 0.01.

RESULTS

The mean values of the physical seed dimensions and percentage oil content of the nine accessions of sesame are shown in Table 2. The seeds of the nine accessions varied significantly (p<0.05) for all the physical dimensions. Accession means showed that seed length varied between 2.65 and 3.77 mm with IBS having the highest value while C-K2 had the least value. Accession IBS had the highest seed width of 2.84 mm and this was about double the value recorded in accession C-K2 (1.49). Seed thickness also varied between 0.61 and 1.28 mm with ALO having the highest value while C-K2 had the lowest value.

Table 2:	Means and standard errors for seed physical dimensions and oil content in nine accessions of sesame
†Means followed by the same alphabets in the same row are not significantly different at p = 0.05level using Duncan multiple range test; *Ranges of values are shown in parenthesis

The 1000 seed weight ranged from 1.52 to 8.42 g with 69B-882 and S530 being the heaviest while C-K2 was the lightest. Length-width ratio ranged from 1.33 in accession IBS to 2.03 in 69B-882 while other accessions were intermediate between the two. Accession C-K2 had the highest length-thickness ratio (5.11) almost twice that recorded for ALO which had the lowest value (2.80). High variation among the accessions was observed for width-thickness ratio with accession IBS having the highest value of 2.98 while the lowest value was recorded for 69B-882 (1.57). Accession ALO had the highest seed geometric diameter (3.91) which was more than quadruple the value recorded for C-K2 (0.81). Seed sphericity ranged from 0.13 in C-K2 to 0.26 in ALO. The greatest variation was observed in seed surface area which ranged from 2.43 in C-K2 to 48.66 mm² in ALO. Seed oil content varied from 53.23 in S530 to 58.56 in IBS. Each accession significantly differed from every other accession.

Seed oil content (SOC) in Table 3 showed positive correlation with seed length, seed width, length-thickness ratio, width-thickness ratio but not significant (p>0.01). In addition, seed oil content recorded negative but non-significant correlation with seed thickness, 1000-seed weight, length-width ratio and seed sphericity. Seed length had positive and significant correlation with seed width (r = 0.874), seed thickness (r = 0.705), seed geometric diameter (r = 0.931), seed sphericity (r = 0.692) and surface area (r = 0.875). Seed width showed positive and significant correlation with seed geometric diameter (r = 0.872) and seed surface area (r = 0.862) however, its correlation with length-width ratio (r = -0.873) was negative and significant. Positive and significant correlation was recorded between seed geometric diameter and characters such as seed sphericity (r = 0.826) and seed surface area (r = 0.984) while the correlation between seed geometric diameter and 1000-seed weight (r = 0.060) was also positive but not significant.

The principal component analysis for seed physical dimensions (Table 4) revealed that the first two principal components accounted for approximately 88.96% of the total variation among the accessions. Most of the genetic variation was explained by the first principal component (56.23%), followed by the second (32.73%). The first component had high positive loadings from seed geometric diameter (0.382), seed sphericity (0.380), seed thickness (0.374), seed surface area (0.362) and seed length (0.354) and high negative loading from the length-thickness ratio (-0.361).

Table 3:	Correlation coefficients of seeds physical dimensions in nine accessions of sesame studied
**Correlation is significant at p<0.01 level. *Correlation is significant at p<0.05 level. ns: Not significant, SLT: Seed length, SWD: Seed width, STH: Seed thickness, SWG: Seed weight, LWR: Length-width ratio, LTR: Length-thichness ratio, WTR: Width-thickness ratio, SGD: Seed geometric diameter, SSP: Seed sphericity, SSA: Seed surface area, SOC: Seed oil content

Table 4:	Eigenvectors and percentage explained variation by the first five principal components of seed physical dimensions of nine accessions of sesame studied

Fig. 2:	The ordination of nine accessions of sesame on Principal Component Axes (PCA) 1 and 2 from cluster analysis of seed physical dimensions

The second component had high positive loadings from width-thickness ratio (0.429), seed oil content (0.381) and seed width (0.362) and high negative loading from 1000-seed weight (-0.430) and length-width ratio (-0.425). However, none of the variables were redundant.

Using the principal components axes 1 and 2, the nine accessions are separated into five groups (Fig. 2). Accession IBS forms the first cluster, EVA and ALO are grouped together in the second cluster, PACH, S530 and 69B-882 form the third cluster, AYK and 65-8B form the fourth cluster while C-K2 forms the fifth cluster. A dendrogram constructed using minimum dissimilarity distance and UPGMA clustering method (Fig. 3) again discriminate among the nine accessions by dividing them into two major genetic groups (Clusters A and B). The genetic dissimilarity coefficients between accessions varied widely, ranging from 0.132 to 0.830. Two major clusters and four subgroups were revealed according to their relationship for seed dimensions and seed oil percentage.

Fig. 3:	Dendrogram generated based on UPGMA clustering method depicting genetic relationship among nine accessions of sesame using seed physical dimensions

Cluster A was subdivided into two subgroups such as 1 and 2 and out of these, only subgroup 1 consisted of sub-subgroups ‘a’ and ‘b’. Subgroup 2 contained only one accession C-K2. This accession had the least values for the seed physical dimensions measured except for length-thickness ratio (5.1) which was the highest. Sub-subgroup 1a contained AYK and 65-8B accessions that were 78% similar and were closely related. Sub-subgroup 1b contained PACH and 69B-882 that were 87% similar and were closely related to S530 with a genetic distance of 0.15. These accessions in sub-subgroup 1b had the highest 1000-seed weight (7.9 g).

Cluster B was subdivided into two subgroups 1 and 2. Out of these, only subgroup 2 consisted of sub-subgroups ‘a’ and ‘b’ while subgroup 1 contained only accession IBS. Accessions ALO and EVA were 82% similar and were separated from IBS with a genetic distance of 0.53. IBS in subgroup 1 had the highest seed length (3.77 mm), seed width (2.84 mm), width-thickness ratio (2.98) and seed oil percentage (58%). ALO and EVA accessions in subgroup2 had the highest seed surface area (47 mm²).

DISCUSSION

Physical seed dimensions such as length, width and thickness are useful in determining aperture size in the design of grain handling machinery. Adebowale et al. (2011) reported that the knowledge of the physical seed dimensions can enhance the selection of the size of perforations for optimum performance of the cleaner as too big or too small a hole may either result in unclean seeds or lesser efficiency of the machine. The range of values obtained for these parameters are within the range for black cumin seeds (Al-Mahasneh et al., 2008), millet (Baryeh, 2002) and local variety of sesame seeds (Tunde-Akintunde and Akintunde, 2004). The range of values for seed geometric diameter obtained in this study (0.81-3.91mm) were lower than that of millet (Baryeh, 2002) and sorghum (Simonyan et al., 2007) but wider than those reported by Adebowale et al. (2011) for sesame seeds.

The degree of sphericity ranged between 0.13 and 0.26 which is lower than the values obtained by Tunde-Akintunde and Akintunde (2004, 2007) and Adebowale et al. (2011) for sesame. The observed differences may be associated with variations in the genotypic background of the sesame seeds employed. However, higher degree of seed sphericity was reported for millet (Baryeh, 2002). The seed surface area in this study ranged from 2.43-48.66 mm² which is wider than those reported by Tunde-Akintunde and Akintunde (2004, 2007) and Adebowale et al. (2011) for both local and improved varieties of sesame. The forgoing implied that the information on physical seed dimension of sesame here could be used to complement already existing information in designing machine for its processing. Furthermore, accession ALO with the highest seed geometric diameter could be incorporated into breeding programme for introgression of this trait into heavy-seeded accessions like S530 and 69B-882 for sesame grain size improvement.

Positive and significant correlations observed between seed length and seed width, thickness, geometric diameter, sphericity and surface area indicate that improvement in one of these attributes will positively affect others in sesame and will consequently lead to increase in seed size. However, non significant correlation between seed oil content and the physical parameters determined suggest that increase in seed size may not necessarily lead to an increase in seed oil content. In principal component analysis, the sign of the loading indicates the direction of the relationship between the component and the variable. In this study, Seed geometric diameter, seed sphericity, seed thickness, seed surface area, seed length and seed width were loaded with positive signs (0.382, 0.380, 0.374, 0.362, 0.354 and 0.287, respectively) and exhibited high correlations. Selection based on these characters may be more efficient in screening for large seed size in sesame. This finding is in line with that of Biabani and Pakniyat (2008) on sesame who further suggested that traits with positive loading and high correlation might be influenced with the same gene or genes.

The clustering patterns using principal component analysis and dendrogram generated by UPGMA of seed physical dimensions also depict genetic relationship among the nine accessions of sesame and both are essentially similar. Accessions IBS, ALO and EVA that clustered together had large seed size based on their high values for the three physical dimensions (i.e. length, width and thickness), seed geometric diameter and surface area. Accessions S530, PACH and 69B-882 were heavy seeded with high 1000-seed weight. The foregoing demonstrates genetic diversity among the nine accessions for the traits determined. Divergent accessions may have good breeding values and accessions in the same cluster may represent members of one heterotic group. According to Genet et al. (2005), the maximum variability for segregation in the segregating population may be achieved by utilizing accessions from different clusters as parents of crosses. Therefore, this information can be used in the selection of parents for crosses in order to maximize expression of heterosis for large seed size in sesame.

CONCLUSION

The results of this investigation has suggested high genetic variability for physical dimensions in sesame. Information from seed size characters could be used to complement already existing data in the development of hopper and dehuller designs for sesame seeds. Selection based on seed geometric dimension, seed sphericity, seed thickness, seed surface area and seed length may be more efficient for suitable sesame genotype screening for large seed size. Furthermore, parents from distinct clusters may be selected in planning crosses for seed size improvement.