Genetic Variability of Eggplant Germplasm Evaluated under Open Field and Glasshouse Cropping Conditions

: Knowledge of agro-morphological genetic variation and cropping conditions on vegetative and yield-related traits plays a significant role in varietal improvement and production of eggplant ( Solanum melongena L.). Following this premise, the current study was conducted to critically asses the genetic variation of 29 eggplant accessions by using agro-morphological characterization evaluated under two cropping conditions, namely, glasshouse and open field. The experiments were laid out in randomized complete block design (RCBD) with three replications. Data on vegetative and yield characteristics were collected and subjected to analysis of variance (ANOVA) using SAS 9.4, while variance components were estimated manually. The results obtained from the analysis of variance indicated a highly significant difference ( p ≤ 0.01) for all characteristics studied in both cropping conditions. The evaluated accessions were grouped into six major clusters based on agro-morphological traits using Unweighted Pair Group Method with Arithmetic mean (UPGMA) dendrogram. Hence, crosses between group I with VI or V could be used to attain higher heterosis and vigor among the accessions. Also, this evaluation could be used as a selection criterion for important yield agronomic traits in eggplant. The methodology and the approaches used may provide a model for the enhancement of other vegetable crop diversity towards adaptability to the cropping condition decision. This result displayed importance for preserving eggplant germplasm for future varietal development and revealed that open field cropping condition is more suitable under Malaysia's agroecology.


Introduction
Eggplant (Solanum melongena L.) is one of the important vegetables belonging to the family Solanaceae, which comprises other significant crop species including chilli (Capsicum annuum L.), tomato (Solanum lycopersicum L.), tobacco (Nicotiana tabacum L.), and potato (Solanum tuberosum L.).

Planting Materials and Agronomic Practice
Twenty-nine eggplants accessions, which form three main populations from Malaysia, Thailand, and China were used in this study, as presented in Table 1. The evaluation was conducted at Field 10 (S8) at Universiti Putra Malaysia (UPM), Serdang, Selangor, Malaysia, which is geographically located between 2°59′ north latitude and 101°42′ east longitude, with 45 m above sea level altitude. The other experiment was conducted in Fertigation Unit, Ladang 15, Universiti Putra Malaysia (UPM), Serdang, Selangor, Malaysia, which is between 2°59′ north latitude and 101°43′ east longitude, with an altitude of 55 m. The experiment was laid out in randomized complete block design (RCBD) with three replications. The environment is hot humid tropics with high humidity and adequate rainfall throughout the year. The seeds were sown in 104-holes seed germinating trays with 1-2 seeds per cell on peat moss growing medium. After 27 days of sowing, the seedlings were transplanted to mixed soil and peat moss with a ratio of 2:1 in polybags for the hardening phase for another 25 days before being transferred to the field and the glasshouse conditions. In each replication, the eggplant was planted with 50 cm spacing between the five plants of each accession and 1 m between the rows. Following the standard cultural practices, agronomic routines and plant maintenance such as fertilizer application, pest and disease management, and weeding were carried out. On a daily basis for the glasshouse fertigation system, the plants were supplied with modified copper formulation fertilizer consisting of N (200 mg L −1 ), P (60 mg L −1 ), K (300 mg L −1 ), Ca (170 mg L −1 ), Mg (50 mg L −1 ), Fe (12 mg L −1 ), Mn (2 mg L −1 ), B (1.5 mg L −1 ), Zn (0.1 mg L −1 ), Cu (0.1 mg L −1 ), and Mo (0.2 mg L −1 ) [15] while the electron conductivity (EC) reading increased in succession according to the growing phase (0.5-3.0). In the aspect of pest and disease mitigation strategies, several pesticides were applied as recommended by the Department of Agriculture, Malaysia (http://jpn.penang.gov.my/index.php/perkhidmatan/teknologi-tanaman/sayur-sayuran/78-terungsp-3424).

Data Collection
Thirteen sets of agro-morphological data were collected and measured from the two planting conditions following the description of the International Boards for Plant Genetic Resources [16] and European Eggplant Genetic Resources Network [17]. To obtain the means of the variables in each plot, five fruits were chosen at random from each of the three tagged plants in the middle of each replication during the harvest. The harvest was carried out five times subject to the productivity of plants that might vary between the accessions. Harvesting frequency and respective number of fruits taken were recorded (data not shown). Plants were evaluated on the yield components fruit girth (FGI), diameter of fruit (DFR), fruit length (FLE), fruit length to width ratio (FLW), total number of fruit (TNF), number of fruit per bunch (NFB), average fruit weight (AFW), and fruit yield per plant (FYP). These also included vegetative parameters such as number of primary branches (NPB), plant height (PHE), stem diameter (SDM), plant spread (PSP), and days to first flowering (DFF). All data measurement and observations were accomplished on the same day to reduce variation in plant growth developmental stage or environmental changes.

Statistical Analysis
All vegetative, yield, and yield-related data in collection were subjected to analysis of variance (ANOVA) using the SAS version 9.4 (SAS Institute Inc., Cary, NC, USA), while means comparisons were separated with least significant difference (LSD) at 5% level of significance. Additionally, grand mean, standard deviation, and coefficient of variation (CV) were recorded for each trait measured. Together with the analysis of variance (ANOVA), nested design analysis was a hierarchical design plotted with fruit type and the whole set of attributes of eggplant as an interest of evaluation in eggplant accession. Among the 29 accessions subjected in this agro-morphological analysis, ten accessions were the round type of eggplant, and 19 accessions were long shape type. Both long and round fruit types of specifically assigned eggplants were subsampled under accession and eventually made the fruit type within accession. Accession has a higher level as compared to fruit type.
Genetic relationships among the eggplant germplasm were determined using the unweighted pair group method with arithmetic mean (UPGMA algorithm) and sequential agglomerative hierarchical non-overlapping (SAHN) methods. Cluster trees [18] are important multivariate tools to assess genetic variation among the eggplant's germplasm under two cropping conditions. Utilizing the basis of comparable vegetative and yield components traits, the grouping of individual accessions was exposed by cluster analysis according to similarity and relatedness of eggplants. Other than that, restricted maximum likelihood (REML) using PROC VARCOMP in SAS version 9.4 (SAS Institute Inc., Cary, NC, USA) was used for variance components estimations. Genetic parameters such as genotypic coefficient variance (GCV), phenotypic coefficient variance (PCV), heritability in the broad sense (H 2 B), and genetic advance (GA) were calculated using the following equation [19].
where K is a constant that represents the selection intensity. At the value when k is at 5%, the rate is 2.06. is the phenotypic standard deviation, and ℎ is the broad-sense heritability value. GA values of 0-10%, 10-20%, and >20% are low, intermediate, and high, respectively [21,24].

Yield and Yield Components Across Two Cropping Conditions
The pooled analyses of variance for yield and yield traits from two cropping conditions are presented in Table 2. Highly significant differences (p ≤ 0.01) were observed among the accessions and the fruit types within the accessions for all yield and yield parameters measured. Similarly, highly significant differences (p ≤ 0.01) were recorded in the cropping conditions for all yield and yield-related traits except for the number of fruits per bunch (NFB). Next, highly significant differences (p ≤ 0.01) were observed in fruit type for all the yield and the yield-related parameters except for average fruit weight (AFW) and fruit yield per plant (FYP), where non-significant differences were observed. On the other hand, the interaction between cropping conditions with the accession showed highly significant differences (p ≤ 0.01) for the most important yield and yield component traits, which were fruit yield per plant (FYP), fruit length (FLE), total number of fruit (TNF), and average fruit weight (AFW). This strongly signified that cropping conditions and agronomic practices play pivotal roles in affecting eggplant varietal aspects in terms of yield and yield components. The result also revealed that there was no sign of replication effect within the cropping condition except for total number of fruits (TNF), average fruit weight (AFW), and fruit yield per plant (FYP). A high coefficient of variation (CV) of more than 40% was applied to average fruit weight (AFW) and fruit yield per plant (FYP), indicating they were the most diverse quantitative agro-morphological characteristics observed in Table 2. In general, all accessions were strongly varied from each other in terms of yield characteristics. The reason for the significant difference is apparently the differences in their origins that cause the existence of variation in a population [25]. Similarly, several studies have been carried out on phenotypic variation among eggplant accessions. The outcome of this research is in agreement with the findings of Caguiat and Hautea [26]. Hence, this strongly supported a postulation by Naujeer [27] that enhanced yield and improved fruit quality are defined as the main objectives in eggplant breeding program.
The yield traits' least significant difference (LSD) mean performances of 29 accessions are presented in Table 3. The fruit girth (FGI) ranged from 22.37 cm to 5.71 cm with an average value of 12.84 cm. Accession 13MR (Malaysian Round) had the longest fruit girth (FGI), while accession 6TL (Thailand Long) showed the smallest fruit girth (FGI). Concurrently, the same accession pattern was observed for the diameter of fruit (DFR) that showed an average diameter of 4.36 cm. Accession 6TL recorded the smallest diameter of 1.99 cm, and the widest diameter of fruit (DFR) was observed in 13MR at 7.31 cm. The fruit length (FLE) varied from 2.10 cm to 17.70 cm. The longest fruit length (FLE) was observed in 2TL, while accession 5TR (Thailand Round) had the shortest fruit length (FLE). The mean fruit length (FLE) among accessions was 10.31 cm. For fruit length to width ratio (FLW), accession 1TR recorded the lowest with ratio (0.73), while a ratio of 5.15 was observed in 10ML (Malaysian Long) with an average mean of 2.65. The highest total number of fruits (TNF) was produced by 15ML (63.17), and accession 4TR produced the lowest number of fruits (TNF) at 4.50 fruits. The average total number of fruits (TNF) produced was 23.19 among the accessions. The average number of fruits per bunch (NFB) was 1.07 with 6TL producing the highest (2.67), and the remaining accession had the lowest number of fruits per bunch (NFB) (1.00) except for 26CL (China Long), which had an intermediate (1.40) total number of fruits per bunch (NFB). The average fruit weight (AFW) was 433.92 g. The fruit weight ranged from 142.50 g (5TR) to 962.90 g (21ML). The overall yield per plant (FYP) means was 1603.36 g. The yield per plant (FYP) ranged from 323.90 g (4TR) to 2932.20 g (13MR). Generally, accession 13 MR had the best performance in fruit yield per plant (FYP) in both cropping conditions with slightly lower yield in glasshouse cropping conditions. Generally, mean comparisons of accession performances between two cropping condition portrayed a higher mean value (as indicated in bold) in the open field except for the number of fruits per bunch (NFB). Indeed, the yield is evergreen major parameters for evaluating cropping conditions. Pollination in the open field is more frequent, as it is aided with natural pollinators such as bees and wind flow to help dissemination and distribution of pollen. Meanwhile, in glasshouse conditions, limited aeration and higher temperature due to cladding materials in this microclimate [28] eventually can reduce the fruit set. Alternatively, hand pollination assistance through shaking flowers is seen as the savior for promoting the set of the first blossoms of the flowers, and this implicitly causes more labor work and requires higher costs. Other than that, eggplant is more susceptible to whiteflies family species (Aleyrodidae sp.) in a glasshouse compared to the open field. This might be due to higher temperatures causing an outbreak of whiteflies due to a thermal tolerance up to 40-45 °C in the glasshouse [29].   open field, GH = glasshouse, FGI = fruit girth (cm), DFR = diameter of fruit (cm), FLE = fruits length (cm), FLW = fruits length to width (ratio), LSD = least significant difference, SEM = standard error of mean, n.s = not significant at p ˃0.05 and means with the same letter in each column also not significantly different at 5% probability level. open field, GH = glasshouse, TNF = total no of fruits (no.), NFB = number of fruits per bunch (no.), AFW = average fruits weight (g), FYP = fruit yield per plant (g), LSD = least significant difference, SEM = standard error of mean, n. s=not significant at p ˃ 0.05 and means with the same letter in each column also not significantly different at 5% probability level.

Vegetative Traits Across Two Cropping Conditions
The combined analysis of variance for vegetative traits is presented in Table 4. Highly significant differences (p ≤ 0.01) were observed among the accessions for number of primary branches (NPB), plant height (PHE), stem diameter (SDM), and days to first flowering (DFF), while non-significant differences were observed in plant spread (PSP). To emphasize, plant height is the most critical vegetative indicator of high yield, as postulated by [25]. Together with other vegetative parameters, these traits contributed by genetic makeup were implicitly influenced by the environment, especially cropping conditions. This was mainly due to limited sources of photosynthates partitioning to meet vigorous sink competition. Hence, the quota for yield was unfairly used by somatic cells growth, which resulted in luxurious vegetative development and obvious height. For fruit type, all vegetative parameters showed no significant differences except for the number of primary branches (NPB), where a highly significant difference was observed, and plant spread (PSP), which showed a significant difference at p ≤ 0.05. The fruit type within accessions showed highly significant differences for all vegetative parameters except for plant spread (PSP), which illustrated no significant difference. While cropping conditions also indicated high significance (p ≤ 0.01) for number of primary branches (NPB), plant height (PHE) and plant spread (PSP) were vegetative components that had an impact on genetic variation; stem diameter (SDM) and days to first flowering (DFF) indicated no significant difference. On the other hand, the interaction between cropping conditions with the accession showed a highly significant difference (p ≤ 0.01) for the stem diameter (SDM) trait only, and the remaining number of primary branches (NPB), plant height (PHE), days to first flowering (DFF), and plant spread (PSP) eventually showed no significant difference. Moreover, the result implied that there was a highly significant difference (p ≤ 0.01) in replication effect within the cropping conditions, which were number of primary branches (NPB), plant height (PHE), and plant spread (PSP). However, days to first flowering (DFF) and stem diameter (SDM) showed no significant difference. had the smallest with 82.81 cm, and 3TR recorded the largest plant spread at 116.14 cm. The average plant spread (PSP) length was 96.50 cm. For days to flowering (DFF), 3TR recorded the longest days to flowering with 100.67 days, while 35CL recorded the earliest at 71.17 days. The average mean of days to flowering (DFF) was 88.81 days. Given the resulting comparison of accession mean performance for two cropping conditions, as shown in Table 5, the means for all vegetative traits were comparatively higher in the greenhouse compared to the open field cropping conditions except for day for first flowering (DFF). This showed robust growth of vegetative yield in the greenhouse cropping condition that may be affected by a significant and continuous supply of fertilizer using irrigation. The variation of vegetative growth among eggplant accessions was wide enough to indicate the perspective glass view for improving accessions studied for all characteristics that eventually support and prepare the reproductive phase of eggplant. It was evidenced that this might be due to the association of genetic composition together with the environment factor applied. Table 5. Means for vegetative characteristics studied in 29 accessions of eggplant across two cropping conditions. open field, GH = glasshouse, NPB = number of primary branches (no.), PHE = plant height 90 days after transplant (cm), PSP = plant spread 90 days after transplant (cm), LSD = least significant difference, SEM = standard error of mean, n.s = not significant at p ˃ 0.05 and means with the same letter in each column also not significantly different at 5% probability level.

Heritability and Genetic Parameters
Broad-sense heritability, phenotypic coefficient variation, genotypic coefficient variation, and genetic advance are presented in Table 6. Heritability is a dimension of physical appearance (phenotypic traits) or total variance that is handed down from the parents [25]. We could identify a bigger range of low to high broad-sense heritabilities observed for most of the yield component traits, while low broad-sense heritabilities were evidenced for all vegetative traits. Estimation of broadsense heritability showed the highest value for the trait fruit girth (FGI) with 77.50% and the lowest for plant spread (PSP) with 0.00. Number of fruits per bunch (NFB) and fruit length gave the values of 74.98% and 63.84% heritability estimations, respectively. Moderate values (30-60%) were observed in fruit length to width ratio (FLW) and diameter of fruit (DFR), while the lowest heritability values were illustrated in number of primary branches (NPB), plant height (PHE), stem diameter (SDM), plant spread (PSP), days to first flowering (DFF), total number of fruits (TNF), average fruit weight (AFW), and fruit yield per plant (FYP). In general, for high heritability (>60%), values together with high genetic advance (>20%) were observed for fruit girth (FGI), fruit length (FLE), and number of fruits per bunch (NFB). These parameters are mainly controlled by the additive type of genes and can be used as selection criteria for significant improvement in fruit yield production of eggplant. The results obtained are in agreement with previous research [30][31][32]. Nevertheless, moderate heritability values but high genetic advance were observed in diameter of fruit (DFR) and fruit length to width ratio (FLW). Both lower heritability values and genetic advance were respectively indicated by plant height (PHE), stem diameter (SDM), plant spread (PSP), and days to first flowering (DFF). This explanation of the function of non-additive genes in the traits could be corrected by heterosis breeding [33,34].
Next, estimation of the phenotypic coefficient of variance (PCV) and the genotypic coefficient of variance (GCV) ranged from zero to 47.34%, portrayed by plant spread (PSP) and fruit length to width ratio (FLW), respectively. As in overall characteristics of vegetative and yield and their components traits, the highest GCV (>20%) was evidenced by FGI (30.92%), FLW (47.34%), FLE (43.85%), NFB (28.72%), and AFW (30.30%), while moderate GCV (10-20%) was observed in NPB (11.69%), SDM (10.70%), DFR (28.18%), TNF (14.97%), and FYP (12.69%). The lowest GCV (<10%) was indicated by the remaining PHE (6.15%), PSP (0.00%), and DFF (6.92%), in which their phenotypic expressions were strongly affected by the environment. Hence, the limited selection was found on these traits. Ranges for PCV values from moderate to high were observed as 15.08% in plant height (PHE) to 80.89% in the total number of fruit (TNF). High PCV was indicated in TNF (80.89%) followed by FYP (79.  [32,[35][36][37][38][39]. Nonetheless, the difference between them proposing the governance of genetics and hence the selection on a phenotypic basis would remain reliable as influenced by environmental factors. This also implicitly shows the importance of germplasm adaptive capacity with the environment used in upcoming crop breeding selection. Higher genotypic coefficient of variation together with high heritability and high genetic advance provide superior indication rather than individual parameters [40]. Fruit girth (FGI), fruit length (FLE), and number of fruits per bunch (NFB) were highlighted trait candidates in meeting these criteria. Thus, it is pivotal to select one trait that gives positive manipulation to the other traits. In addition to the performance response in both vegetative and yield parameters, this study also highlighted the considerably significant degree of genetic variation with the evidence among the accession for some traits that could be further explored for the breeding program. This finding is also in agreement with other researchers, such as [35,41,42]. Additionally, accession selection would be resourceful for an eggplant improvement program established based on yield and yield component traits in particular.

Cluster Analysis of Agro-Morphological Traits
Agro-morphological characteristics including vegetative, yield, and yield components parameters were adopted based on the Euclidean distances among the 29 accessions of eggplant to construct a UPGMA dendrogram as in Figure 1. This dendrogram revealed 29 eggplant accessions clustered into six groups with a similarity coefficient of 0.35 and which were the best fit for convenience discussion, and this implied a high level of agro-morphological variation of eggplant accessions. As presented in Table 7, cluster I had four admixed accessions from Thailand and China,  namely 1TR, 26CL, 5TR, and 6TL, while cluster II had the largest group of 19 accessions (2TL, 27CL,  20ML, 34CL, 10ML, 22ML, 3TR, 17ML, 18ML, 13MR, 14ML, 16ML, 29MN, 32MN, 7TR, 25ML, 19ML, 15ML, and 30MN). Meanwhile, cluster III had 9ML, 23ML, and 35CL-three admixed accessions from Malaysia and China-and the remaining clusters, IV, V, and VI, had one accession each, 8ML, 21ML, and 4TR, respectively (two accessions from Malaysia and the latter from Thailand). Indeed, there are diverse eggplant accessions commercially domesticated between these three origins, and agromorphological traits are reliable to classify different accessions in a pool of germplasm [25,43]. Regarding morphological traits' mean performances between clusters, as shown in Table 8, cluster II achieved the best in terms of yield and yield component traits. It portrayed a fruit yield per plant of 1867.95 g/plant, which was mainly due to the highest total number of fruits (TNF) of 26.47. Utilizing the basis of comparable vegetative and yield components traits, the grouping of individual accessions was exposed by cluster analysis according to similarity and relatedness of eggplants. Hence, the large difference of each accession attribute performance proposed in the crosses between group I and IV or V could be used to attain higher heterosis and vigor among the accessions.

Conclusions and Recommendation
This research revealed that eggplant germplasm had ample genetic variation portrayed through agro-morphological characterization via ANOVA and multivariate analysis. The pivotal understanding of agro-morphological evaluation of genetic variation on eggplant germplasm synergized with cropping condition practices leads to the finding of higher eggplant production with preferable cropping conditions. In Malaysia, it was found that the open field is more suitable for eggplant production with better efficiency of agronomic management together with sustainable production systems. Nevertheless, future work is suggested to explain the molecular approach of genetic variation together with a comprehensive validation of a few seasonal and site trials.