Rice, genetic divergence, D2 statistics, molecular diversity, gene-specific markers.

Rice (Oryza sativa L.) is an important cereal and staple food crop for more than half of the population in the world. Over the world, Asia alone produces 90% of the world's rice and consumed by over 2 billion people to derive 80% of their energy needs and it contains 80% carbohydrates, 7-8% protein, 3% fat and 3% fibre (Chaudhari et al., 2018).

Rice is encoded with phenomenal genetic diversity also with hundreds and thousands of germplasms in gene banks. India, as a part of the centre of origin, rice is endowed with a wide variety of germplasm, which includes landraces, obsolete cultivars, wild or weedy species, improved varieties, and so on, resulting in immense genetic diversity. The genus Oryza comprising of 24 species, of them only 2 species such as Oryza sativa and Oryza glaberrima are cultivated in Asia and Africa, respectively. The most cultivated species Oryza sativa is again sub-divided into indica, japonica, javanica, aus (deep-water rice) and aromatic (Basmati, Jasmine, Joha etc.) (Garris et al., 2005). 

The knowledge of genetic divergence is very important in the selection of suitable parents for hybridization under varying conditions. The wider the diversity among the parents, the greater will be the chance of obtaining heterotic combinations. A hybridization programme combining genetically varied parents from distinct clusters would give a chance to bring together gene constellations of heterogeneous type, with promising hybrid descendants resulting from the complimentary interaction of divergent genes in parents. 

Conventionally, genetic diversity can be estimated by means of D² analysis, developed by Mahalanobis (1936), which is a technique based on multivariate analysis and is found to be an efficient tool in quantifying the degree of divergence in germplasm and determining the relative contribution of each character towards total divergence. 

However, the selection of parents based on phenotype is more effective and precise when it is combined with selection based on genotype since the phenotype also includes environmental factors. Diversity in genes of rice accessions creates an option for the breeder to select desirable traits and use them in making new combinations (Garris et al., 2005). At the genotypic level, one of the best methods for analysing diversity in rice is by molecular markers which can tell huge differences among accessions at the DNA level, providing a more reliable and well-planned aid in accession characterization and genetic make-up.

The recent improvements and advances in high throughput sequencing techniques facilitated high-quality reference genomes, which led to numerous molecular markers, high-density genetic/physical maps, several QTLs and candidate genes for key traits. Further, several gene-specific markers have been developed and deployed for the improvement of several rice popular varieties like BPT5204, Pusa Basmati, Swarna, Tellahamsa, Tetep, etc. Thus, the present study aimed the assessment of genetic diversity at both morphological and molecular levels for yield and yield-attributing traits among the rice genotypes. This would help in the identifying diverse parents in the population for the development of new varieties having heterotic combinations.

The present experiment was carried out using 40 rice genotypes during Rabi, 2022 at Wetland farm, Sri Venkateswara Agricultural College, Tirupati. Each genotype was grown in two rows of two meters length with a spacing of 20 cm between rows 15 cm between plants in a Randomized Block Design with three replications. All the recommended cultural practices were followed for raising a good and healthy crop. The data was recorded on 14 yield and yield attributing traits viz., days to 50% flowering, days to maturity, plant height (cm), number of tillers plant-1, number of panicles plant-1, panicle length (cm), total number of grains panicle-1, number of chaffy grains panicle-1, number of filled grains panicle-1, grain length (mm), grain width (mm), length/width ratio of rice grain, 1000 grain weight (g) and grain yield plant-1 (g). Observations were collected on five randomly selected plants from each genotype in each replication, with the exception of days to 50% flowering and days to maturity, which were recorded on a plot basis. The recorded data was subjected to analysis of variance and Mahalanobis D2 statistics were used for genetic divergence analysis. The genotypes were clustered by using Tocher's method. The intra- and inter-cluster distances were calculated and were used to describe the genotype relationship with the help of the formula proposed by Singh et al. (2016). To perform the statistical analysis INDOSTAT software was used.

Table 1: List of 40 rice genotypes and their pedigree.

DNA isolation and PCR assay. The genomic DNA was isolated from leaves of 20-25 days old seedlings of 40 rice genotypes using Cetyl Trimethyl Ammonium Bromide (CTAB) method (Murray and Thompson 1980). The isolated DNA was quantified using Nano drop spectrophotometer (ND-1000, Thermo Scientific, Nanodrop Technologies, U.S.A). 10µl of PCR mixture consists of 1µl of PCR buffer with Mg2+, 0.5 µl of dNTPs, forward and reverse primers of 0.5 µl each, 0.1 µl of Taq DNA polymerase, 5.4 µl of autoclaved double distilled water and 2 µl of template DNA. Amplification was done using an Eppendorf thermo cycler, with the temperature profiles of initial denaturation at 94oC for 5 min, denaturation at 94 oC for 30 sec, annealing (55-65 oC) based on primer for 30 sec, extension at 72 oC for 1.0 min, final extension for 10 min at 72oC for 35 cycles and storage at 4 oC for 10-15 min. The amplified PCR products were electrophoresed on a 3% agarose gel stained with ethidium bromide (10mg/ml) at 100 volts for 1.5 hours in 1X TBE buffer. A 50 bp or 100 bp ladder (Genei) was used to determine the optimal product size. The gel pictures were taken under UV light with the Syngene Ingenius geldoc system.

Table 2: List of gene-specific markers related to yield included in the study.

Molecular data analysis. The amplified products for marker analysis were scored visually based on the presence (taken as ‘1’) or absence (taken as ‘0’) of band for each primer. Each marker fragment was treated as a unit character and only clear and unambiguous bands were scored. Genetic diversity parameters like number of alleles per locus, major allele frequency and heterozygosity were estimated by using markers data. The allele frequency represents the frequency of a particular allele for each marker; while heterozygosity is the proportion of heterozygous individuals in the population. Polymorphic information content (PIC) was estimated using the following formula (Hwang et al., 2009; Barik et al., 2019): 

where i = 1 to n and Pij is the frequency of jth allele for the ith band scored for a particular marker.

Molecular diversity analysis for the genotypes was according to unweighted neighbour joining method and a dendrogram was constructed using Jaccard’s dissimilarity coefficient in DARwin software version 6.0.21 (Perrier, 2006).

Morphological diversity by using D2 statistics

Estimation of D2 Values. The mean values of 40 genotypes [(X1)-(X2)] were converted into standardized uncorrelated mean values [(Y1)-(Y2)] using the pivotal condensation technique. D2 values were calculated for all possible [40 (40-1)/2] 780 pairs of genotypes.

Grouping of Genotypes into Clusters. All the 40 genotypes were clustered into eight distinct clusters using Tocher’s method (Rao, 1952). The genotypes present within the cluster had smaller D2 values than those with different clusters. Table 3 and Fig. 1 shows how genotypes are distributed into distinct clusters. Out of eight clusters, clusters I was the largest one comprising of 17 genotypes followed by cluster III with 13 genotypes, clusters II and IV had 3 genotypes each. Clusters V, VI, VII and VIII are monogenotypic clusters. The genotypes present in the monogenotypic clusters were unique and very useful for breeding purpose.

Table 3: Distribution of 40 rice genotypes into eight clusters based on Tocher’s method.

Average Intra and Inter Cluster D2 Values. The average intra and inter cluster D2 values among the eight clusters were represented in Table 4 and the cluster diagram was furnished in Fig. 2. The average intra-cluster distance ranged from 0.00 to 219.32. Maximum intra-cluster distance was reported in cluster III (219.32) followed by cluster II (147.28), cluster I (133.82) and cluster IV (105.22) which manifested that some divergence still existed among the genotypes of the same cluster, which could be used in the yield improvement whereas no intra-cluster distance was observed in clusters V, VI, VII and VIII because of monogenic clusters.

The inter-cluster D2 values ranged from 231.47 to 1853.65. The maximum inter-cluster distance (1853.65) was recorded between cluster VI and VIII followed by cluster VII and VIII (1278.73), cluster III and IV (1246.12) and cluster V and VIII (1179.41). It may be extrapolated that the genotypes in these clusters have the maximum genetic diversity. Hence, the genotypes from these clusters could be utilized in the crossing programme to develop promising genotypes. Whereas, minimum inter-cluster distance of 231.47 was recorded between cluster V and VII, followed by cluster I and II (277.38), cluster I and V (279.94) and cluster VI and VII (296.07) which denoted that genotypes of these clusters were genetically close. Inter-cluster distances were greater than intra-cluster distances, indicating that there is more genetic diversity between clusters rather than within cluster performance.

Table 4: Average Inter (above diagonal) and Intra (diagonal) cluster distances (D2 values) for eight clusters of 40 rice genotypes.

Fig. 1. Cluster diagram representing inter and intra-cluster distances among eight clusters of 40 rice genotypes.

Mahalanobis Euclidean Distance (Not to the Scale)

Cluster Means for Yield and Yield Attributes. The cluster means for 14 yield and yield attributing traits were presented in Table 5. Considerable variation among the cluster means for all the characters indicated the divergent nature of clusters formed.

Table 5: Cluster means for yield and yield attributing traits among 40 rice genotypes.

DFF: Days to 50% flowering; DM:Days to maturity; PH: Plant height (cm); TP: Number of tillers plant-1; PP: Number of panicles plant-1; PL: Panicle length (cm); TG: Total number of grains panicle-1; CG: Number of chaffy grains panicle-1; FG: Number of filled grains   panicle-1; GL: Grain length (mm); GW: Grain width (mm); LWR: Length/width ratio of rice grain; TGW: 1000 grain weight (g); GYP: Grain yield plant-1 (g)

Early flowering was recorded in the genotypes of cluster IV (89.67 days), while late flowering was recorded in the genotypes of cluster VIII (113.64 days) with the general mean of 102.62 days. Cluster I, II, IV, V and VI had lower values than the cluster mean. Days to maturity ranged from 120.67 days in cluster IV to 145.18 days in cluster III and cluster I, cluster II, cluster IV, cluster V and cluster VI recorded lower cluster means for days to maturity than the general mean (133.57 days).

The genotypes of cluster VI were shorter in plant height (70.63 cm), while genotypes of cluster VII were taller in plant height (96.47 cm). The clusters that recorded lower values than the general mean (85.81 cm) for plant height were cluster I, II, IV, VI and VIII. Cluster mean for number of tillers plant-1 was highest in cluster IV (16.40) and lowest in cluster VII (9.73), higher cluster mean than the general mean (11.79) was recorded in clusters IV and VI. Cluster means for number of panicles plant-1 was highest in cluster VI (12.27) and lowest in cluster VII (8.60), cluster means that exceeded the general mean (10.01) were clusters IV, V and VI. Cluster means for panicle length ranged from 20.37 cm (cluster VI) to 27.03 cm (cluster VII). The superior clusters for panicle length that exceeded the general mean (22.89 cm) were clusters V and VII.

The cluster means for total number of grains panicle-1 varied from 101.42 (cluster IV) to 290.87 (cluster VII). The clusters III, VI and VII recorded a higher number of grains panicle-1 than the general mean (197.47). Number of chaffy grains panicle-1 exhibited an overall mean value of 18.53 with cluster means ranging from 5.76 (cluster IV) to 39.40 (cluster VII), while the clusters that were below the general mean value were I, II, IV, V and VIII. The cluster means for number of filled grains panicle-1 varied from 95.67 (cluster IV) to 251.47 (cluster VII). The clusters III, VI and VII recorded a higher number of filled grains panicle-1 than the general mean (178.95).

The maximum and minimum cluster means observed for grain length were 9.60 mm in cluster V and 7.50 mm in cluster III, respectively. The general mean of 8.39 mm was exceeded by the clusters II, V, VII and VIII. The cluster mean for grain width was ranged from 2.17 mm in cluster VII to 3.19 mm in cluster VIII, cluster means that were lower than the general mean (2.60 mm) was recorded in clusters I, III, V, VI and VII. The cluster mean for length/width ratio of rice grain was ranged from 2.46 in cluster IV to 4.08 in cluster VII, the cluster means that were superior than the general mean (3.29) was recorded in clusters II, V, VI and VII.

The cluster means for 1000 grain weight ranged from 12.52 g (cluster VI) to 25.22 g (cluster VIII), higher values than the general mean for 1000 grain weight (18.31 g) were recorded in clusters II, IV and VIII. Cluster means for grain yield plant-1 ranged from 23.44 g (cluster IV) to 37.30g (cluster VI) and the clusters III, VI, VII and VIII were superior with higher values than the general mean of 29.29 g.

Cluster VI comprised of RNR19186 recorded desirable values for plant height, number of panicles plant-1 and grain yield plant-1.Cluster V comprised of DRR Dhan 38 recorded desirable value for grain length. Cluster VII comprised of WGL1142 recorded desirable values for panicle length, grain width, length/width ratio of rice grain, total number of grains panicle-1 and number of filled grains panicle-1. Cluster IV registered a desirable value for days to 50% flowering, days to maturity, number of tillers plant-1 and number of chaffy grains panicle-1. Cluster VIII consisted MTU3626 which recorded desirable value for 1000 grain weight. Hence it is inferred that crossing between the genotypes from these clusters could be advised to generate a wide spectrum of variability followed by effective selection for these characters in later generations.

Relative Contribution of Each Character towards Diversity. The number of times each of the 14 characters appeared in first rank and its respective percent contribution towards diversity was presented in Table 6.

Table 6: Percent contribution of 14 yield and yield attributing traits towards genetic divergence in 40 rice genotypes.

Among all the characters studied, maximum contribution towards genetic divergence was recorded by days to 50% flowering (48.08%) by obtaining the first rank which was followed by grain length (16.92%), 1000 grain weight (16.79%), grain width (9.87%), length/width ratio of rice grain (2.44%), number of filled grains panicle-1 (2.18%) and grain yield plant-1 (1.28%). The characters viz., total number of grains panicle-1 (0.90%), plant height (0.64%), panicle length (0.64%), number of panicles plant-1 (0.13%) and number of chaffy grains panicle-1 (0.13%) contributed least towards genetic divergence.

In the present study, days to 50% flowering, grain length and 1000 grain weight were found the best discriminatory characters for selection of diverse genotypes. So, these characters could be exploited maximum to get varieties with a higher yield.

The results with respect to the relative contribution of each character towards diversity were supported by earlier findings of Vanlalrinngama et al. (2023); Lakshmi et al. (2022); Roy et al. (2022); Singhal et al. (2022); Naik et al. (2021); Chandramohan et al. (2016)  for days to 50% flowering; Devi et al. (2022); Sujitha et al. (2020) ; Singh et al. (2008) for grain length; Gayathri et al. (2023); Lakshmi et al. (2022); Singhal et al. (2022); Guru et al. (2017); Chandramohan et al. (2016) for 1000 grain weight. 

Molecular diversity by using UPGMA method

Molecular Marker Analysis. A total of 13 gene-specific markers related to yield (Table 2) were used for the assessment of molecular diversity among 40 rice genotypes. These 13 gene-specific markers were spread on seven chromosomes 2, 3, 4, 5, 7, 8 and 9 of rice genome. Out of 13 markers, seven markers viz., Dep1-S7, Dep1-S9, RM7289, PAY1SP6, RGS-1, RMw513 and Spike-Indel3 produced clear polymorphic amplicons in 40 rice genotypes and the remaining six markers HY2-4, RM16942, RID-711, RM5499, RM21945 and RM502 were found to be monomorphic. A total of 15 alleles were identified with seven gene-specific markers across 40 rice genotypes. The number of alleles per locus observed maximum of three (Spike-Indel3) and the remaining markers with two alleles each with an average of 2.143 alleles per locus. The Polymorphic Information Content (PIC) values for the used gene-specific markers ranged from 0.134 to 0.524 with an average of 0.289. The highest PIC value was obtained for the marker RGS-1 (0.524) followed by Dep1-S7 (0.414), PAY1SP6 (0.375), Spike-Indel3 (0.296), RM7289 (0.139), RMw513 (0.139) and Dep1-S9 (0.134).

Fig. 2. Gel picture showing the allelic pattern in 40 rice genotypes with Dep1-S7.

Table 7: Number of alleles, allele size and PIC value of the gene-specific markers.

Markers with PIC values of 0.5 or above are very informative for genetic investigations and especially effective in differentiating the polymorphism rate of a marker at a specific locus (Virk et al., 1995). In this study, marker with PIC value ≥ 0.5 is only one i.e., RGS-1 indicating this marker has higher discriminating power when compared to other markers.

As we have selected gene-specific markers, that are most favoured by breeders during their selection. The number of alleles and PIC values of these gene-specific markers were significantly lower than the frequently utilized SSR markers from prior studies. For example, Choudhary et al. (2013) found an average of 3.6 alleles per locus with a PIC of 0.87 using 52 SSR markers in 100 rice genotypes, whereas Vigneshwari et al. (2017) found an average of 4.2 alleles per locus with a PIC of 0.453 using 15 SSR markers. However, in accordance with the current study, Ngangkham et al. (2018) reported a total of 21 alleles with an average of 2.1 allele per locus and a PIC value ranging from 0.13 to 0.58 with an average value of 0.31 using 10 gene-specific markers regulating grain size in 89 rice genotypes.

Molecular Diversity Analysis. The data from gene-specific markers related to yield was used to analyse the genetic diversity among the 40 rice genotypes. The dissimilarity matrix was estimated using seven gene-specific markers based on Jaccard’s dissimilarity coefficient using DARwin version 6.0.21 (Perrier, 2006). The calculated dissimilarity matrix was used for the clustering of 40 rice genotypes. An un-weighted neighbour-joining (UNJ) method was used for the dendrogram construction with 1000 permutations to determine the bootstrap values. The dendrogram was presented in Fig. 3.

Fig. 3. Dendrogram of 40 rice genotypes based on molecular data of gene-specific markers using Jaccard’s dissimilarity coefficient by UPGMA method.

Table 8: Grouping of 40 rice genotypes into three clusters by using gene-specific markers data based on Jaccard’s dissimilarity coefficient by UPGMA method.

The results led to the clustering of 40 rice genotypes into three clusters, cluster I, cluster II and cluster III. Among the three clusters, cluster I is the largest comprising of 25 genotypes followed by cluster II with 9 genotypes and cluster III with 6 genotypes. Genotypes viz., Warangal samba, RNR19186 and 187-3 in Cluster I were having one common parent BPT5204; MTU1061 and MTU1210 were derived from one common parent MTU1001; ND3 and ND44 were derived from one common parent NLR34449; MTU1318 and LRG-5 were grouped in cluster I along with their parents MTU7029 and NLR33892, respectively. The remaining genotypes viz., Pant dhan 12, Sharbati, MM152, Navara, Anjali, Heera, BPT1235, NLR34242, Aditya, Udayagiri and NLR2422 were grouped under the cluster I may be due to the common ancestral origin. In Cluster II, one genotype i.e., WGL1142 along with its parent WGL32100 grouped together. Out of the six genotypes in cluster III, two genotypes i.e., MTU1153 and MD5 along with their common parent, MTU1010 were clustered together.

Upadhyay et al. (2012) have previously documented genotype grouping based on the effect of the pedigree, which is consistent with the current study, categorized the 25 popular rice varieties into two major clusters and showed that varieties with at least one common parent were grouped in one cluster; Choudhary et al. (2013) showed that varieties released during different decades were also grouped together due to the presence of common parents in their pedigree; Singh et al. (2016) showed that varieties sharing common parentage were grouped in the same cluster; and Vigneshwari et al. (2017) showed the relationship between the clustering pattern of the thirteen rice varieties was obviously due to the pedigree. There were a few cases where genotypes with shared parentage were not grouped in the same cluster. The two mutants from MTU1010, MM11 and MM152, fell into distinct groups.

V. Bhanu Prakash, P. Shanthi, M. Shanthi Priya, P.V. Ramana Rao, P. Lavanya Kumari, M. Reddi Sekhar, A. Manasa, G. Anil Kumar, K. Girish Kumar and V.L.N. Reddy  (2024). Molecular and Morphological Genetic Diversity for Yield and Yield Attributing Traits in Rice (Oryza sativa L.). Biological Forum – An International Journal, 16(3): 166-174.