Introduction

The African eggplant (Solanum aethiopicum L.) is one of the important indigenous fruit vegetables widely grown and consumed across most regions of tropical Africa. It is the third most consumed fruit vegetable after tomato, pepper and onion both in quantity and value in the region (Osei, Banful, Osei, & Oluoch, 2010). Mature fruits of African eggplant are eaten fresh, with fried groundnuts or used to prepare special delicacies called 'African salad' in southern Nigeria (Igwe, Akunyili, & Ogbogu, 2003). A significant increase has been observed in its production across sub-Saharan Africa from 606,672 tonnes in 1994 to 2,079,920 tonnes in 2018 (FAO, IFAD, UNICEF, WFP and WHO, 2018).

Eggplant is considered amongst the healthiest fruit vegetables for its low calories and high concentration of various macro and micro minerals essential for maintaining good health (Docimo et al., 2016). They are rich sources of fibres, vitamins (A, B1, B2, B6, B12, C, D), magnesium, calcium and iron even though potassium is the most abundant mineral ranging from 200 to 600mg/100g of fresh matter (Kowalski, Kowalska, & Wiercinski, 2003; Nimenibo & Omotayo, 2019; Nyadanu & Lowor, 2015). The crop has been reported to play an essential role in meeting the nutritional needs of the Igbo-speaking tribe in southern Nigeria where consumption of fresh fruits might be of great benefit to glaucoma patients and to prevent heart disease (Denkyirah, 2013; Igwe et al., 2003). S. aethiopicum is used in the management and treatment of diarrhoea and hypertension (Adeniji & Aloyce, 2012). The high yield and nutritive value of the leaves and fruits complemented with resistance to pests and diseases endear the crop to consumers, farmers and researchers (Bonsu, Owusu, Nkansah, Oppong-Konadu, & Adu-Dapaah, 1998; Taher, Rakha, Ramasamy, Solberg, & Schafleitner, 2019; Toppino, Vale, & Rotino, 2008).

Eggplants belong to the Solanaceae family, which encompasses three closely related cultivated species endemic to Afro-Eurasia. Two sections exist at the subgenus level, namely Melongena and Oliganthes sections. The section Melongena comprises two species (S. melongena and S. macrocarpon) while the Oliganthes group has only one species (S. aethiopicum). S. aethiopicum has been grouped into four different ecotypes or cultivars including Aculetum, Gilo, Kumba and Shum groups as revealed by similarities in genotypic characterization through varied phenotypes (Sharmin, Khalil, Begum, & Meah, 2011). Aculetum is mostly used as ornamental, Gilo is used for its fruits, Kumba is for both fruits and leaves while Shum is used for its leaves (Lester & Daunay, 2003).

Consumer preferences for an African eggplant cultivar are based on a number of traits including fruit size, form, fruit colour and taste (sweet or bitter). Morphological characterization using conventional descriptors has proved useful for describing and establishing relationships among cultivar groups and accessions in scarlet eggplants (Adeniji, Kusolwa, & Reuben, 2013). The enormous morphological variability present in the eggplant family, despite being characterized by a narrow genetic base, might be attributed to new segregants emanating from natural hybridization and backcrossing (Meyer, Karol, Little, Nee, & Litt, 2012).

Despite their socioeconomic significance and their role in meeting the nutritional needs of the ever-increasing population across sub-Saharan Africa, these heirloom and indigenous adapted cultivars are becoming less popular, and the efforts to improve them for traits of interest to farmers and end-users are scarce (Bationo-Kando et al., 2015).

Continuous planting and selection of many diverse cultivars of S. aethiopicum by small-scale farmers as well as the existence of germplasm collections have helped to conserve the majority of desired traits within families over the years. The long period of selection by these poorly resourced farmers has resulted in a number of landraces exhibiting different variants with unique traits such as earliness, colour, size and taste. In essence, the African eggplant has long been neglected by formal crop improvement programmes except in breeding programmes where it is used as a source of specific traits. Furthermore, they are considered neglected and underutilized crops since their nutritional and economic potentials are mostly underexploited (Padulosi, Phrang, & Rosado-May, 2019).

Systematic characterization of African eggplant accessions using morphological and nutritional traits is an important prerequisite toward their conservation and use in further studies and genetic improvement in the region (AVRDC, 2003). Unfortunately, minimal efforts have been directed to identify and select promising genotypes with a good combination of desired agronomic and nutritional qualities that could be used as parental materials for hybridization or released as new open-pollinated varieties. The process of germplasm characterization and trait screening to identify and select desired gene combinations can be challenging.

This study characterized newly selected and purified eggplant accessions collected in Nigeria to ascertain their mineral composition and to identify promising elite lines with the best combination of desired agronomic traits that could further be deployed for eggplant genetic improvement programmes.

Materials and methods

Results

The frequency distribution of qualitative traits observed in all 20 accessions is presented in Table 2. All the genotypes (100%) exhibited green stems, petioles and sepals. At the reproductive stage, 48.48% of the fruits had a white colour, 45.45% expressed lemon green while 3.03% were light green and 3.03% exhibited deep green fruit colour. Three prominent fruit shapes were observed: oval (51.51%), long (30.30%) and round (18.18%). All accessions exhibited perpendicular fruit position with 69.69% and 30.30% of the populations having pointed and flat ends respectively. The fruits of selected eggplant accessions are presented in Figure 1.

Table 3 lists the descriptive statistics measures of spread: mean, range, standard deviation and coefficient of variation (CV). The partitioning of the means revealed high significant variations for all traits at P ≤ 0.01. For most traits, higher variations in terms of range and CV were observed in the nutritional data compared to the phenotypic data. The highest CV was recorded for calcium (43.91%) followed by average number of fruits per plant (37.13%) and average yield per plant (32.21%). The top performers for selected yield-contributing traits and nutritional parameters are NHEPA54, NHEPA39-1, NHEPA10, NHEPA10, NHEPA1, NHEPA56, NHEPA23, for vitamin C, iron, calcium, days to flowering, number of branches, plant height at maturity and number of fruits per plant, respectively, while NHEPA54 was outstanding for high yield potential and vitamin C content. Accession NHEPA10 was the top performer for calcium and days to flowering while average number of fruits per cluster had NHEPA17, NHEPA19, NHEPA23 as the top performers.

Principal component analysis (PCA), which is a statistical technique used to emphasize variation and bring out strong patterns in data sets, was performed to show the traits that best contributed to the observed genetic variation. The eigenvalues and proportion of accounted variance for each variable are shown in Table 4. PC1 had an eigenvalue of 4.027 while PC2, PC3 and PC4 had eigenvalues of 2.889, 1.861 and 1.363, respectively. The first four principal component axes (PCA) accounted for 28.77%, 20.64%, 13.29%, 9.73% of the total variation individually and, cumulatively, 72.42% of the total variability while the first two PCs contributed 49.41% (Figure 2). The first PC axis, which accounted for the highest proportion (28.77%) of the variability, was dominated by traits with relatively high factor scores (˃ 2.60) corresponding to number of branches, plant height at maturity, number of fruits per cluster, number of fruits per plant, and fruit width. The second PC axis was dominated by days to flowering, fruit length, stem diameter, petiole length and pedicel length. Also, the third PC axis was dominated by average yield per plant, fruit length, stem diameter, petiole length and pedicel length while the fourth PC axis was dominated by days to flowering, number of branches, fruit length, fruit width and petiole length.

Correlations between pairs of quantitative variables are recorded in Table 5. There was no significant association between the nutritional parameters except for a negative moderate significant association between vitamin C and iron (r = -0.50, P < 0.05). Iron content correlated positively with number of branches (r = 0.487, P < 0.05), average number of fruits per plant (r = 0.532, P < 0.05) and number of fruits per cluster (r = 0.551, P < 0.05). Plant height had a negative but moderate significant association with iron (r = -0.461, P < 0.05) and calcium (r = -0.407, P < 0.05) but was positively correlated with vitamin C (r = 0.492, P < 0.05). The strongest and most persistent correlation was recorded for association between fruit width and plant height (r = 0.574, P < 0.01) and number of fruits per plant (r = -0.737, P < 0.01); stem diameter with days to flowering (r = 0.734, P < 0.01). A positive significant correlation was observed between pedicel length and fruit length (r = 0.561, P < 0.05), and between pedicel length and petiole length (r = 0.528, P < 0.05).

Based on variation in the phenotypic parameters the 20 eggplant accessions were clustered into four unique groups (Figure 3). Clusters I and II contained seven and three accessions, respectively, while clusters III and IV both had five accessions. Means of variables, ranges and standard deviation for each cluster are presented in Table 6. Cluster III was unique in having accessions with high vitamin C content and high yield potential while clusters II was characterized by accessions with high iron content, an increased number of fruits per plant and a higher number of fruits per cluster. Clusters I and IV were characterized by early maturing accessions dominated by top-performing accessions in fruit-related traits (fruit length and fruit width respectively).

Discussion

The success of genetic improvement programmes in enhancing desired traits of interest to farmers and breeders depends on the magnitude of genetic variability available in the germplasm and the extent to which the desirable traits are heritable. The high significant variation observed for most qualitative and quantitative traits considered in this study establishes the feasibility of imposing selections towards the improvement of desired traits of interest in African eggplant. Frequency distribution among the qualitative traits with a preponderance of fruits characterized by white to cream colours and lemon green suggests that the majority of the accessions belong to the S. aetihiopicum group. This supports earlier reports by (Osei et al., 2010) that eggplant accessions belonging to S. aethiopicum had mixtures involving cream white to light yellow fruits; thus, fruit colour combined with fruit shape might be considered a strong phenotypic marker in characterizing eggplant taxa in Africa.

The high CVs and range for some of the quantitative characters could be attributed to genetic variations from natural crossings and ecogeographical factors. The maximum and minimum mean values could present a rough estimate of the variation in magnitude of variability present among genotypes. Traits such as average number of fruits per plant and fruit width that exhibited a high range of variation had more scope for improvement in the eggplant population.

The principal component analysis identified traits that contributed the most to observed variations within a group of entries (Grittins, 1975; Sneath & Sokal, 1973). The first four principal component axes in the current study accounted for 72.42% of the total variability measured. The first principal component analysis had the highest discriminating ability (contributing 28.77% out of 79.47% of variability from the first four axes) and was dominated by traits with relatively high factor scores (˃ 2.60) corresponding to number of branches, plant height at maturity, number of fruits per cluster, number of fruits per plant, and fruit width. This is in agreement with Clifford and Stephen (1975) who reported that the first principal component axis was the most important in reflecting the variation patterns among accessions and that the characters highly associated with these should be used in differentiating the accessions. Furthermore, this is in line with the findings of (Iezzoni & Pritts, 1991) and (Chikaleke, 2018) who reported that the implication of principal components can be accessed from the contribution of the different variables to each principal component (PC).

Correlation analysis is used to identify the relationship between variables (Anshori et al., 2018) and to facilitate the identification of elite traits to rely on in selection exercises of a breeding programme. The positive significant association between iron content, number of branches, number of fruits per cluster and number of fruit per plant; number of fruit per plant with number of branches and number of fruits per cluster; fruit width and plant height at maturity; stem diameter and days to flowering; pedicel length with fruit length and with petiole length will facilitate selection of eggplant accessions with a good combination of these traits. The significant positive correlation displayed by these traits is in agreement with (Dhaka & Soni, 2013) who reported a positive significant association between yield and yield-related traits. However, where the traits had significant negative correlation coefficients (vitamin C and iron content, plant height at maturity and iron content, plant height at maturity and calcium content, number of fruits per cluster and plant height at maturity, number of fruits per plant and plant height at maturity, fruit width with number of fruits per cluster and with number of fruits per plant) indicates that an increase in one trait might lead to a decrease in the other trait or vice versa. This is in agreement with (Mazer, Delesalle, & Neal, 1999) and (Nyadanu et al., 2015) who reported that traits with significant inverse relationships could be improved independently among eggplant accessions in Ghana. However, selecting tall plants in this eggplant population might result in an indirect selection for low calcium content, while favouring increased iron content will lead to selecting genotypes with low vitamin C. Selection pressure can be deployed for an increased number of fruits per plant to simultaneously increase iron content and number of branches. Similar observations were reported by (Arivalagan, Bhardwaj, Gangopadhyay, Prasad, & Sarkar, 2013) and (Nyadanu et al., 2015) in their earlier works on mineral composition and morphological characterization of eggplant.

The cluster analysis emphasized further the relative contribution of various quantitative parameters to the total variability. The grouping of accessions in each cluster based on quantitative descriptors could be attributed to the fact that these accessions share some similarities. The high-yielding accessions in cluster III (NHEPA54) expressing high vitamin C and iron content could be deployed as progenitors to combine with eggplant genotypes from cluster I and create a new gene combination with improved calcium content and yield potential. Creating new eggplant varieties with high-yield potential and increased vitamins and minerals (iron, calcium and vitamin C) will not only increase farmers’ income but will also help to reduce health challenges associated with hidden hunger among the rural and urban populace in the region. Selection and hybridization of genotypes from clusters I and IV such as NHEPA35 and Yalo will produce new segregants characterized by bigger fruits with increased iron concentration.

Conclusion

This study successfully characterized 20 eggplant accessions for phenotypic and nutritional traits of interest and identified top-performing new eggplant accessions with unique traits that could be deployed in crosses to facilitate the step-wise creation of new eggplant varieties with the best combination of desired traits. Furthermore, selection in favour of yield-increasing traits such as number of fruits per plant and number of fruits per cluster that showed a significant positive correlation with iron, will lead to selecting genotypes with increased iron content and higher yield potential simultaneously. Top-performing accessions for iron (NHEPA39-1), calcium (NHEPA39-1) and vitamin C content (NHEPA54) identified in this study should be deployed for hybridization to create new eggplant varieties with improved nutritional content.

Author contributions

Olawale Olsesan Oguntolu: study design, execution, drafting.

Conflict of interest statement

The authors declare no conflict of interest. All authors approved the final manuscript.