Yield potency assessment and characters association of promising lines of mustard ( Brassica rapa L.) in southern region of Bangladesh

Mustard ( Brassica spp .) is a major oilseed crop and the world’s third largest source of vegetable oil. During the rabi season of 2022–2023, the experiment was carried out at the Regional Agricultural Research Station, Bangladesh Agricultural Research Institute (BARI), Barishal to evaluate the performance of 12 genotypes of Brassica rapa L., including the check variety BARI Sarisha-14. The experiment was conducted under randomized complete block design (RCBD) design with three replications. Analysis of variance demonstrated that a highly significant (p ≤ 0.01) differences for the traits viz., plant height, no. of primary branches, no. of secondary branches, and no. of seeds per siliqua whereas no.


Introduction
Rapeseed-mustard belongs to the family of Cruciferae under the genus Brassica are most important oilseed crop, a source of vegetable oil as well as a widely expanding oilseed crop in Bangladesh [1] .There are about ten distinct oil crops in Bangladesh that produce fats and oils of varying quality and quantity.Among them, rapes (Brassica campestris L. and Brassica napus L.) and mustard (Brassica juncea L.) are key sources of vegetable fats.In Bangladesh, the most widely cultivated species are Brassica rapa and Brassica juncea.Both the plant species that are grown in Bangladesh, Brassica juncea and Brassica campestris, are together referred to as "mustard" [1,2] .
In Bangladesh, the usage of edible oil is rising daily.Bangladesh's agricultural sector includes mustard production and cultivation as essential components.Since ancient times, Bangladesh has cultivated mustard, as a significant edible oilseed crop [3] .Mustard seeds provide 40-45% oil, 20-25% protein, and 12-18% carbohydrates [4] .Mustard is high in fat-soluble vitamins such as A, D, E, and K, as well as energy (approximately 9 kcalg −1 ) [5] .Its oil not only serves as an excellent fat alternative in our daily diet but also stimulates the nation's economy [6] .
The area devoted to growing mustard has grown dramatically in recent years.Examining the current domestic mustard output in Bangladesh and projecting the future is crucial for fulfilling the rising demand.Bangladesh produced 11.64 lakh tones of mustard oil seed in the current fiscal year of 2022-2023 against a target of 11.12 tones due to a rise in acreage as well as a high yield [7] .Despite the production rate increasing in terms of our demand, oil production is still not remarkable.As a result, a considerable volume of oil and oil seeds are imported into our country each year.Increasing mustard production might be one strategy to reduce the country's edible oil demand-supply mismatch [8] .Bangladesh Agricultural Research Institute (BARI) has created a variety of high-yielding mustard varieties with yield potential of up to 2.5-2.8t ha −1 in recent years, as well as launched some promising lines in terms of global climate change [9] .Although these new types have a lot of potentials, there are several difficulties that make it tough to produce the optimum results.If these problems could be solved, we could become self-sufficient in the production of edible oil.To fit into the pre-existing cropping pattern, the majority of farmers in our nation grow the traditional variety Tori-7.When compared to other recently established contemporary varieties, the yield of the native Tori-7 is incredibly low.Consequently, the yield potential of newly developed promising lines is very high.In light of the foregoing, the present experiment was undertaken under regional conditions to find out early lines of Brassica rapa L. with better agronomic traits for yield and wider adaptability.

Materials and methods
The experiment was carried out at Rahmatpur, Barishal, during Rabi 2022-2023, with 12 genotypes of Brassica rapa L. with yellow seed coat color and one control, BARI Sarisha-14.The experiment was designed with three replications in a randomized complete block design.The plot measured 3 m × 0.9 m.Continuous seed planting in rows 30 cm apart was carried out at Rahmatpur on November 18, 2022.After a few days of germination, the seedlings were trimmed to 5 cm apart.Fertilizers were administered at a rate of 120:80:60:40:4:1 kg ha -1 of N:P:K:S:Zn and Boron derived from Urea, TSP, MOP, Gypsum, Zinc Sulphate, and Boric acid [9] .During the last stage of land preparation, half of the urea and all other fertilizers were applied.
The remaining urea was sprayed at the bloom initiation stage.To ensure a satisfactory yield, all intercultural operations were completed on schedule.Days to 50% flowering, days to maturity, plant height (cm), number of primary and secondary branches per plant, number of siliqua per plant, number of seeds per siliqua, 1000 seed weight (g), and seed yield per plot were all recorded.The yield of the plot was converted to kilograms per hectare (kg ha −1 ).All the data was statistically examined.The meteorological data during the cropping season is listed in Table 1.

Statistical analysis
For all of the morphological features presented, an analysis of variance was performed with Statistical Analysis System (SAS) version 9.4.The significant differences were separated using List Significant Difference (LSD) at 5%.According to Khalikqi et al. [10] , Pearson's correlation was utilized to determine the associations between yield and yield component qualities using "proc corr" in the SAS software.The Euclidian Distance Method, as well as Dices' and Jaccard's similarity of coefficient data, were used to investigate genetic diversity.In addition, the genetic inter-relationship (showing dendrogram) among Brassica rapa L. genotypes was estimated using the Unweighted Pair Group Method using Arithmetic Average (UPGMA) and the algorithm & sequential, agglomerative, hierarchic, and non-overlapping (SAHN) method.For this study, NTSYSpc version 2.1 (Numerical Taxonomy Multivariate Study System), Exeter Software, Setauket, NY, USA software 4.0 was used.Principal component analysis (PCA) was performed using comparable software to generate the two-dimensional (2D) plots described by Khan et al. [11] .

Mean square (MS) of analysis of variance (ANOVA)
Table 2 shows the results of an analysis of variance (ANOVA) for nine characteristics from 12 genotypes and one control.Among the genotypes, highly significant (p ≤ 0.01) mean square accounted for plant height (PH), number of primary branches per plant (NPBPP), number of secondary branches per plant (NSBPP), and number of seeds per siliqua (NSPS) whereas no. of siliqua per plant (NSPP), 1000 seeds weight (TSW), yield (kg ha −1 ) had significant difference (p ≤ 0.05).This incidence revealed the presence of highly significant genetic variability among the genotypes in terms of statistically significant traits.However, among the replication only the trait plant height (PH) and number of seeds per siliqua (NSPS) showed significant difference (p ≤ 0.05).Fayyaz and Amin [12] also found that there was significant genetic and environmental variability among the genotypes due to variance in analysis of variance.denotes significance level at p ≤ 0.01; Rep = replication; Gen = genotype; DF = degree of freedom; LSD = least significant difference; D50%F = days to fifty percent flowering; DM = days to maturity; PH = Plant height (cm); NPBPP = no. of primary branch per plant; NSBPP = no. of secondary branch per plant; NSPP = no. of siliqua per plant; NSPS = no. of seeds per siliqua; TSW = 1000 seed weight (g) and seed yield (kg ha -1 ).

Clustering pattern and Principal component analysis
As indicated by a dendrogram, the UPGMA (average linkage) cluster analysis identified distinct clusters showing linkages among tested accessions (Figure 2).With a dissimilarity coefficient of 0.022, the accessions were grouped into four primary clusters based on their evaluated quantitative features.To pick the optimal cluster number and readability, the dendrogram was cut off at 0.022 using Mojena [16] stopping criterion.Cluster I absorbed the most accessions (9) with an average yield of 1526.59 kg ha −1 , accounting for 26.25% of the cluster mean yield.Cluster II (28.91%) had the highest average yield (1681.67 kg ha −1 ) with just one accession (Table 3), followed by Cluster I (1526.59kg ha −1 ; 26.25%) and Cluster IV (1319.33 kg ha −1 ; 22.68%) with the best agronomic features.Cluster III genotypes (1287.67 kg ha −1 ; 22.14%), on the other hand, indicated limited yielding potential.Furthermore, cluster I and II had 1.6% and 11.93% higher (+) mean yields compared to the grand mean yield (1502.33 kg ha −1 ) respectively, whereas cluster III (14.28%) and cluster IV (12.18%) had lower (−) yields.In terms of yield, genotypes in clusters I and II (BC-100614(8)-7) provide greater yields than genotypes in clusters III and IV (Table 4) and were recognized as potential accessions for future crop development.Khan et al. [11] discovered a similar tendency in his studies on Bambara groundnut.

Principal component analysis
According to our findings, the first principal component (PC1) accounted for greater variation (54.86%) than the second main component (PC2) (16.89%), whereas the third principal component (PC3) accounts for 83.17% of all variances (Table 5).The majority of the overall variance is explained by the first axis (PC1) of any principal component analysis (PCA) [17] .The variables such as no. of primary branch per plant, no. of secondary branch per plant, no. of siliqua per plant, no. of seeds per siliqua, 1000 seed weight (g) and seed yield (kg ha −1 ) occupied high values in PC1.These traits were positioned into positive (+ve) quartile and very close to each other (Figure 3) in component pattern plot indicating that there is a significant contribution of these traits on yield.In combination with principal component analysis, cluster analysis investigated the relationships between genotypes in terms of seed yield and associated agronomic parameters [11] .The twodimensional (2D) graphical explication (Figure 3) showed that while a small number of accessions were distributed at large distances, the majority were spread at modest distances as shown by the eigenvector in Table 5.The farthest genotypes from the centroid were BC-100614(8)-7, BC-100614(4)-5, and BC-100614(8)-2whereas other accessions were near to centroid.Khan et al. [17] classified the 44 Bambara groundnut accessions based on quantitative features using PCA analysis.Note: PC = principal component; D50%F = days to fifty percent flowering; DM = days to maturity; PH = Plant height (cm); NPBPP = no. of primary branch per plant; NSBPP = no. of secondary branch per plant; NSPP = no. of siliqua per plant; NSPS = no. of seeds per siliqua; TSW = 1000 seed weight (g) and seed yield (kg ha −1 ).

Characters association analysis
The phenotypic correlation among the 9 quantitative traits of twelve Brassica rapa L. genotypes are given in Table 6.The trait yield revealed positively moderate (0.25 ≤ r ≤ 0.75) and highly significant association with no. of primary branches per plant (r = 0.50; p ≤ 0.01), no. of secondary branches per plant (r = 0.42; p ≤ 0.05), no. of siliqua per plant (r = 0.48; p ≤ 0.05) and thousand seed weight (r = 0.33; p ≤ 0.05).A positively strong (0.75 ≤ r ≤ 1.00) and highly significant association was noted for number of primary branches with secondary branches per plant (r = 0.83; p ≤ 0.01).A positive and highly significant association was noted for number of primary and secondary branches per plant with number of siliqua per plant, number of seeds per siliqua, and 1000 seed weight.Number of seeds per siliqua had positive and highly significant interrelation with 1000 seed weight (r = 0.56; p ≤ 0.01).A weak (0 ≤ r ≤ 0.25) and non-significant association was found with yield and no. of seeds per siliqua (r = 0.22; p ≥ 0.05) whereas the traits days to 50% flowering and days to maturity had negative correlation with yield indicating the factor early maturity reduces the yield potentiality of genotypes.The presence of positive and highly significant correlation coefficients between the yield and yield-related traits reported in this study was a sign that the morphological traits measured correctly predicted the yield and were suitable for genotype selection in subsequent breeding programs to improve heterosis or vigor.According to Khan et al. [18] the positive association also suggests that these features should be better explored in order to create genotypes that are desired.Note: "**" correlation is significant at the 0.01 level; "*" correlation is significant at the 0.05 level; D50%F = days to fifty percent flowering; DM = days to maturity; PH = Plant height (cm); NPBPP = no. of primary branch per plant; NSBPP = no. of secondary branch per plant; NSPP = no. of siliqua per plant; NSPS = no. of seeds per siliqua; TSW = 1000 seed weight (g) and seed yield (kg ha −1 ).

Conclusion
The genotypes BC-100614(8)-7, BC-100614(8)-1, BC-100614(7)-3, BC-100614(4)-10, and BC-100614(4)-11 were chosen for further varietal evaluation under regional yield trial in the following year after being found to be promising in terms of all statistical parameters and traits performance, particularly maturity period, yield, and yield contributing characters.The traits number of primary branches, number of secondary branches, number of siliqua per plant and thousand seed weight showed the statistically meaningful association with yield indicated that these traits are beneficial for the selection of genotypes for future breeding scheme.

Figure 1 .
Figure 1.The distribution of yield (kg per hectare) of 1.2 genotypes of Brassica rapa L.

Figure 2 .
Figure 2. Cluster analysis revealed a dendrogram for 12 Brassica rapa L. genotypes based on the UPGMA method of SAHN clustering.

Figure 3 .
Figure 3. Two-dimensional (2D) graph showing the relationship among 12Brassica rapa L. genotypes using PCA revealed by NTSYSpc and component pattern plot for 9 morphological traits revealed by PCA using SAS.

Table 1 .
Meteorological data during the cropping season.

Table 2 .
Mean square of ANOVA for yield and yield components of Brassica rapa L.

Table 3 .
Mean performance and relative position of Brassica rapa L. genotypes.
Note: Means with the same letter are not significantly different.

Table 5 .
Principal component analysis for nine quantitative traits of Brassica rapa L. genotypes.