Study of divergence in non-basmati rice (Oryza sativa L.) for grain yield during reproductive stages under drought condition

Introduction

Rice is a monocotyledon plant, chromosome number 2n = 24, with ∼ 430 MB genome size. Rice belongs to the Gramineae (Poaceae) grass family in cereals. In tropical, subtropical and warm temperate temperatures all across the world, there are 22 wild and 2 cultivated species in the genus Oryza. The two cultivars are Oryza sativa and Oryza glaberrima. Oryza sativa originated in South East Asia (Vaughan, 1994) [11] while Oryza glaberimma is indigenous to West Africa (Niger River delta). Oryza glaberrima is only grown in a few locations such as West Africa, whereas Oryza sativa is produced all over the world. The Oryza sativa is further divided into the indica, japonica, and javanica subspecies. Short-grained, sticky Japonica rice is generally farmed in dry regions. A non-sticky, long-grained kind of rice called indica is frequently grown in fields that are submerged. Both varieties are available in a broad variety (RABBANI 2010) [9] . Rice is a starchy cereal grain and highly selfpollinating crop with Southeast Asian origins. More over half of the world's population relies on rice as a staple diet, making up more than 20% of all calories consumed. Rice is the second most extensively grown cereal grain in the world. The world's most important food crop is rice (Oryza sativa). The amount of genetic variability present in the population is the primary requirement for developing high yielding rice varieties outperforming present day cultivated varieties. Thus, studies on genetic diversity to know the nature and degree of divergence for yield traits will help in designing efficient breeding programme by facilitating the selection of elite parental lines which can be further used in hybridization programme to develop superior and high yielding varieties. Also cluster analysis offers a best method for selecting parents to realize high heterosis to strengthen hybrid breeding programme in rice. With this background the present study was carried out to understand and classify the magnitude and nature of genetic diversity among the studied germplasm lines.

Materials and Methods

Experimental materials:

The experimental material consists of 49 non-basmati rice lines with broad genetics base along with 4 non-basmati lines as check varieties were used for the study the name of 53 genotypes is given in the table 1.

Formula to calculate Genetic Divergence

The genetic divergence in fifty-three genotypes of rice was estimated using Mahalanobis D2 statistic (1936) following Rao (1952) [10] .

D2 = λ Ij δIδJ

Where, λ Ij = reciprocal matrix to the common dispersion matrix. δI = Difference between the mean values of the two populations for I th character δJ = Difference between the mean values of the two populations for J th character. The parameter D 2 can be estimated as D 2 static (Rao, 1952) [10] D 2 = ∑∑ δIJ, didJ = Sample estimate of XIJ Where, δIJ = Sample estimates of X ij Sample estimates of δI di dJ = Sample estimates of δJ Unstandardized and correlated variables (mean value of character) are transformed into standardized uncorrelated variables. After transformation X1, X2, ……………….Xp variable (unstandardized and correlated) becomes Y1, Y2……………….Yp variable (standardized and uncorrelated variables). This transformation was done according to the square root of Dwyer, described by Rao (1952) [10] . The variable transformations is done because this formula requires the inversion of n order of determinants and then evaluation of (n + 1)/2 term whose sum is D 2 . Inter and intra-cluster distances were calculated by Tocher's method as suggested by Rao (1952) [10] to form the clusters. This criterion was based on the concept that populations belonging to the same cluster should, on average, have lower D 2 value than populations from separate clusters.

Results

Genetic Divergence

D 2 statistics which is the most important and potent biometrical tool in quantifying the degree of divergence in various crop plants was used in the present investigation. D 2 statistics utilizes multivariate distribution of natural variation to construct the dispersion matrix. The analysis of variance revealed significant differences among genotypes for all the characters studied and thereby, indicating existence of a good amount of genetic variability. All the genotypes differed from each other with respect to these characters (Table 2&5). Cluster analysis of genotypes was performed based on all characters using Tocher method and then groups were compared.

Clustering Pattern

Cluster analysis, based on all the studied traits, divided the genotypes into six clusters in control condition and six clusters in stress conditions (Table 4&7). Among the six clusters, cluster V consisted maximum number of genotypes (16) followed by cluster I consist 12 genotypes and cluster II, III, IV all consisted 7 genotypes and VI cluster consist 4 genotypes. whereas in stress condition, among the six clusters, cluster V consisted maximum number of genotypes (14) followed by cluster I consist 12 genotypes and III consist 8 genotypes, cluster II and IV each consist 7 genotypes and cluster VI consist 5 genotypes.

Clustering Distance

In control condition the intra and inter cluster average distances among six clusters were variable according to the result represented in (Table 3). Intra cluster distance ranged from 3.731 to 3.085 and 3.805 to 3.097 in control and stress conditions, respectively. (Table 3 &6) whereas the highest intra-cluster distance was recorded for cluster III (3.731) followed by cluster VI (3.704), cluster V (3.221), cluster II (3.163), cluster IV (3.092), and cluster I (3.025) in control condition. Maximum intra cluster distance was observed in cluster III (3.805) followed by cluster VI (3.463), cluster I (3.427), cluster V (3.214), cluster IV (3.113), cluster II (3.097) in stress condition (Table ) indicating wide genetic variability within the genotypes of these clusters. The highest inter-cluster distance was observed between cluster III and VI (5.392), followed by cluster I and VI (5.267), cluster II and III (4.671), cluster II and V (4.514), cluster II and VI (4.444), cluster III and IV (4.204), cluster V and VI (3.868), cluster III and V (3.79), cluster II and IV (3.77), cluster I and IV (3.737), cluster I and II (3.674), cluster I and III (3.632), cluster I and V (3.282) cluster IV and V (3.053) in control condition, while in stress condition the highest inter-cluster distance was observed between cluster III and VI (6.856), followed by cluster I and VI (6.597), cluster IV and VI (5.186), cluster II and VI (5.087), cluster II and III (4.834), cluster V and VI (4.775), cluster III and IV (4.301), cluster III and V (3.946), cluster I and V (3.816), cluster I and III (3.722), cluster I and II (3.533), cluster I and V (3.435), cluster IV and V (2.798) indicating wider genetic diversity among the genotypes between these groups.

Cluster Mean

A comparison of the mean values of different clusters for nineteen traits has been presented in (Table 2&5). Considerable differences in cluster mean values were evident for all the characters studied. The present study revealed that cluster I had highest mean value for 1000-seed weight (19.90), hulling percentage (77.66), milling percentage (64.91) and grain length (6.89)

Discussion

The importance of identifying and selecting genetically diverse plant materials using discriminant functions was first emphasized by Fisher (1936) [4] . Among various multivariate techniques, Mahalanobis D² analysis has been recognized as a reliable and accurate method for estimating genetic divergence based on multiple quantitative traits. However, its effectiveness depends on the stability of the D² measure under different environmental conditions. Earlier studies suggested that genetic diversity has often been associated with geographical or ecological variation (Joshi and Dhawan, 1966) [6] , although genetic drift, natural selection, and human-mediated selection across different agroecological regions can generate greater diversity than geographical separation alone. Arunachalam (1977) [2] highlighted that genetic divergence is an effective criterion for identifying well-defined populations and selecting diverse parents in crop improvement programmes. Murty and Arunachalam (1966) [8] demonstrated several advantages of D² analysis. The method utilizes a dispersion matrix that reflects the natural variation among traits, distinguishes genotypic and environmental variability through replicated observations, and assigns self-weighted importance to each character based on its contribution to total variation, thereby minimizing selection bias. Furthermore, D² analysis enables simultaneous evaluation of a large number of genotypes and traits with computational efficiency. In the present investigation, 53 rice genotypes were evaluated using Mahalanobis D² statistics based on 16 quantitative traits under both control and reproductive-stage drought stress conditions. The genotypes were grouped into six clusters in each environment, indicating substantial genetic divergence. The clustering pattern confirmed that geographical origin alone is not a reliable indicator of genetic diversity, supporting the usefulness of Mahalanobis D² analysis for identifying genetically diverse parents irrespective of their origin. Under control conditions, Cluster V contained the highest number of genotypes (16), followed by Cluster I (12), Clusters II, III and IV (7 each), and Cluster VI (4). In stress conditions, Cluster V again contained the maximum number of genotypes (14), followed by Cluster I (12), Cluster III (8), Clusters II and IV (7 each), and Cluster VI (5). In both environments, inter-cluster distances were greater than intracluster distances, indicating broad genetic diversity among clusters. The highest intra-cluster distance was recorded in Cluster III (D² = 3.731) under control conditions and Cluster III (D² = 3.805) under stress conditions, indicating greater variability within these clusters.

Cluster mean analysis further revealed substantial variation among clusters for yield and yield-related traits. Under control conditions, Cluster VI recorded the highest biological yield per plot (2630.27) and grain yield per plot (625.17), Cluster III exhibited the highest number of effective tillers per plant (7.98) and grain width (2.44), whereas Cluster V recorded the highest panicle length (21.65) and flag leaf length (25.96). Under Drought stress, Cluster VI showed the highest biological yield per plot (2119.66) and total number of grains per panicle (134.04), Cluster II recorded the highest number of effective tillers per plant (7.14) and days to 50% flowering (134.52), while Cluster V exhibited the highest panicle length (18.16) and flag leaf length (25.09). These findings are consistent with earlier reports by Banumathy et al. (2010) [3] , Arivoli et al. (2009) [1] , and Hosan et al. (2010) [5] .

Conclusion

The genotypes were divided into six clusters under control conditions and stress conditions using Mahalanobis D 2 statistics and the Tocher technique. The highest inter-cluster distance was observed between cluster III and VI, followed by cluster I and VI, cluster II and III, cluster II and V, cluster II and VI, cluster III and IV, cluster V and VI, cluster III and V, cluster II and IV, cluster I and IV, cluster I and II, cluster I and III, cluster I and V, Cluster IV and V in control condition, while in stress condition the highest inter-cluster distance was observed between cluster III and VI, followed by cluster I and VI, cluster IV and VI, cluster II and VI, cluster II and III, cluster V and VI, cluster III and IV, cluster III and V, cluster I and V , cluster I and III, cluster I and II, cluster I and V, cluster IV and V indicating wider genetic diversity among the genotypes between these groups.

References

1. Arivoli V, Saravanan K, Prakash M. A study on D2 analysis in rice. International Journal of Plant Sciences, Muzaffarnagar. 2009;4(1):157-160.
2. Arunachalam V. Heterosis for characters governed by two genes. Journal of Genetics. 1977;63(1):15-24.
3. Banumathy S, Manimaran R, Sheeba A, Manivannan N, Ramya B, Kumar D, et al. Genetic diversity analysis of rice germplasm lines for yield attributing traits. Electronic Journal of Plant Breeding. 2010;1(4):500504.
4. Fisher RA. Statistical methods for research workers. 1936.
5. Hosan SM, Sultana N, Iftekharudduala KM, Ahmed MNU, Mia S. Genetic divergence in landraces of Bangladesh rice (Oryza sativa L.). The Agriculturists. 2010;8(2):28-34.
6. Joshi AB, Dhawan NL. Genetic improvement in yield with special reference to self-fertilizing crops. 1966.
7. Mahalanobis PC. A note on the statistical and biometric writings of Karl Pearson. Sankhyā: The Indian Journal of Statistics (1933-1960). 1936;2(4):411-422.
8. Murty BR, Arunachalam V. The nature of divergence in relation to breeding systems in some crop plants. Indian Journal of Genetics and Plant Breeding. 1966;26:188198.
9. Rabbani MA, Masood MS, Shinwari ZK, Shinozaki KY. Genetic analysis of basmati and non-basmati Pakistani rice (Oryza sativa L.) cultivars using microsatellite markers. Pak J Bot. 2010;42(4):25512564.
10. Rao CR. Advanced statistical methods in biometric research. 1952.
11. Vaughan DA. The wild relatives of rice: a genetic resources handbook. 1994.