Genetic diversity assessment of common winter wheat (Triticum aestivum L.) genotypes

The knowledge about of genetic diversity of common winter wheat (Triticum aestivum L.) genotypes is useful for production of more efficient crops adapted to diverse conditions. A set of 32 common winter wheat varieties was used to estimate and then utilize the genetic diversity between common winter wheat genotypes by using cluster analysis and factor analysis and to identify effective factors on genetic improvement. Eight agronomic traits were included in the study. The GCV values were lower than PCV values for all the traits. High genetic advance combined with high heritability showed characters: plant height and spike length. Cluster analysis based on ward's method using Euclidian distance, grouped the cultivars into four clusters. Genotypes in the second group were in the highest rate with respect to number of productive tillers per plant, number of spikelets per spike, number of grains per spike, grain weight per spike and yield per plant. Principal Component analysis was applied to group accessions according to similarity on the basis of eight traits in tree components in the factor plane. The first three components explained 82.63% of total variation in the experiment. Cluster analysis based on the three factors grouped the varieties into four clusters. Genotypes of the first cluster can be used for increase in number of productive tillers per plant in breeding programs. Genotypes of the second cluster can be used for increase in thousand grain weight. Genotypes in the fourth cluster had highest mean with respect to first factor and can be used for increase in number of grain per spike.


Introduction
The knowledge of genetic diversity and relatedness in the germplasm is a pre-requisite for crop improvement programmes.Decreases of genetic variation within cultivars as result of conventional breeding practices affect the success for productivity and adaptability improvement of crops -particularly bread and durum wheat.Because approaching to the limits of biological productivity as result of global climate change the need of new genetic pools is essential (Stoeva et al., 2009).Breeding for developing of accessions with high yield and grain quality is a long and complex process, the success of which is by no means guaranteed (Tsenov et al., 2009).So precise information on the nature and degree of genetic diversity present in wheat collections from its principal areas of cultivation would help to select parents for evolving superior varieties (Islam, 2004;Khodadadi et al., 2011;Aharizad et al., 2012).For the genetic amelioration of this crop, diverse genotypes from the existing germplasm should be selected and used in further breeding programme (Shashikala, 2006).There are various analyses and statistical approaches for genetic diversity identification -cluster analysis, PCA and factor analysis (Bhatt, 1970;Carves et al., 1987;Mohammadi and Prasanna, 2003;Eivazi et al., 2007;Khodadadi et al., 2011).Hierarchical cluster analysis can be used to estimate genetic dissimilarity and similarity in collections.According to Peeters and Martinelli (1989) the analysis could have applications for the selection of parental lines for which varying degrees of segregation are sought.Grouping genotypes at the basis of studied characteristics is one of the suitable methods for determining nearness, distance and closeness of them.Factor analysis has been used by many researchers to determine the factors with contributed to the variation of quantitative traits in common winter wheat.
The aim of this study was to estimate and then utilize the potential genetic diversity between common winter wheat genotypes and to identify effective factors on genetic improvement.

Materials and Methods
The study is carried out in the experimental field of Institute of Plant Genetic Resources "Konstantin Malkov"-Sadovo, Bulgaria during 2011-2013 growing seasons.Thirty two common winter wheat varieties from different countries (Rumania, Serbia, Hungary and China) were examined (Table 1).The cultivars were sown in harvest plots each of 10 m2 in three replications, in the randomized block design.Normal agronomic and cultural practices were applied to the experiment throughout the growing seasons.The agronomic characters were taken after harvesting the plants.From each variety, 20 plants were collected for biometrical measurements.Data were registered for number of productive tillers per plant, plant height, spike length, number of spikelets per spike, number of grains per spike, grain weight per spike, thousand grain weight and grain yield per plant.The mean data from all eight characters were used to analysis of variance according to Lidansky (1988).Genotypic and phenotypic variances, genotypic and phenotypic coefficient of variability, broad sense heritability were calculated according to Singh & Chaudhary (1985).Genetic advance in terms of percentage of means was assessed by Brim et al. (1959).Hierarchical cluster analysis (Ward's method) using Euclidian distance was performed by using the STATISTICA software.PC-analysis was applied to group accessions according to similarity on the basis of eight traits (number of productive tillers per plant, plant height, spike length, number of spikelets per spike, number of grains per spike, grain weight per spike, thousand grain weight and grain yield per plant) in three components in the factor plane by using SPSS 13 software and the related clusters were plotted based on the main components.

Analysis of variance
The results of analysis of variance are shown in Table 2.The values of mean squares for all eight characters indicated highly significant differences between common winter wheat accessions (Table 2).
Phenotypic variance ranged from 0.42 (grain weight per spike) to 201.09 (plant height).Values of genotypic variances ranged between 0.08 (grain weight per spike) and 166.95 (plant height) (Table 3).Higher phenotypic and genotypic variance values of 201.09 and 166.95 for plant height and 142.35 and 40.39 for number of grains per spike were recorded in the study.The Genotypic coefficient of variability (GCV) ranged from 5.23% for number of spikelets per spike to 27.66% for grain yield per plant, whereas, phenotypic coefficient of variability (PCV) ranged from 8.69% for number of spikelets per spike to 51.00% for grain yield per plant (Table 3).The recorded of GCV were high for grain yield per plant (27.66%), moderate for number of productive tillers per plant (19.18%), spike length (18.28%), plant height (14.05%), grain weight per spike (13.14%) and number of grains per spike (11.79%).The characters: number of spikelets per spike and thousand grain weight recorded low GCV estimates.The recorded of PCV were high for grain yield per plant (51.00%), grain weight per spike (29.92%), number of productive tillers per plant (27.83%), number of grains per spike (22.13%) and spike length (20.08%), moderate for thousand grain weight (17.78%) and plant height (15.42%).

Estimation of broad sense heritability and genetic advance
Moderate heritability was recorded for number of productive tillers per plant (47.51%) and number of spikelets per spike (36.28%).High heritability estimates were observed for plant height (83.02 %) and spike length (82.87%).Genetic advance expressed as percentage of mean was high for spike length (34.28%), grain weight per spike (30.90%), number of productive tillers per plant (27.28%) and plant height (26.37%).

Clustering of genotypes
Hierarchical cluster analysis based on Euclidean dissimilarity using Ward's method classified the tested wheat cultivars into four significant different clusters at 50% linkage distance (Figure 1).The first cluster included six genotypes, the second cluster seven, the third cluster six and the forth cluster thirteen genotypes.Analysis of mean and standard deviation were made for all four clusters (Table 4).Feng kang 8, Da mai zi, Long chun 8, Long chun 5, Zapashna and Harkivska 81 were grouped in the first cluster including 18.75% of total genotypes.The average values of genotypes in this cluster for number of productive tillers per plant, plant height, grain weight per spike, thousand grain weight and grain yield per plant were higher than the total means of all genotypes (Table 4).Standard deviations of traits in this group except number of productive tillers per plant, number of grains per spike and grain weight per spike were less than the total standard deviations.Genotypes in this cluster had greatest values for plant height (110.00) and 1000 grain weight (48.55).Rusiya, Lillyana, Renesansa, Gora, Evropa 90, Pobeda and Lyutizha genotypes were classified in second cluster including 21.87% of total genotypes.Genotypes in this group were in the highest rate with respect to number of productive tillers per plant (6.26), number of spikelets per spike (20.86), number of grains per spike (58.17), grain weight per spike (2.37) and yield per plant (9.73).Karma, Izvor, Dragana, Rapsodia, Golosa and MV-Lepeny genotypes were classified in third cluster including 18.75% of total genotypes.The average values of traits in this group except plant height were less than total means of all genotypes.In the fourth cluster 13 genotypes were classified including 40.62% of total varieties.Values of number of productive tillers per plant, plant height, thousand grain weight and grain yield per plant in this cluster were less than the total means of all genotypes.Genotypes in this cluster were in the highest rate with respect to spike length (14.69) and in the least rate with respect to plant height.

Principal component analysis (PC-analysis)
PC-analysis was applied to arrange accessions by their similarity.The values of the three components to each of the studied parameters were calculated empirically.The first three factors explained 82.63% of total variation (Table 5).First factor determining 36.72% of the variation has an important role to justify alteration of spike length, number of spikelets per spike, number of grains per spike and grain per spike (Table 5).Second factor had justified 29.91% of total variance.The factorial coefficients of number of productive tillers per plant and grain yield per plant were high and positive.In the third factor that had explained 16.00% of alteration, thousand grain weight had the greatest effect.
Cluster analysis based on the three factors grouped the varieties into four groups (Fig. 2).Average of factors for each cluster is shown in table 6.In the first cluster, five accessions were classified including 15.62% of total genotypes.Varieties in this cluster were in the highest rate with respect to second factors.Second group comprises 9 varieties including 28.13% of total genotypes.The varieties in this cluster had greatest values for third factor.Therefore, these genotypes were superior with respect to 1000 grain weight.In the third group, 11 genotypes were classified including 34.38% of total accessions.In the fourth group, 7 varieties were classified 21.87% of total genotypes.Genotypes in this cluster had highest mean with respect to first factor.

Discussion
In present study, 32 genotypes of common winter wheat varieties were studied to assess their genetic potential.The analysis of variance showed the presence of significant differences between common winter wheat accessions for all traits analyzed with probability F< p 0.001.According to many researcher as Kamat (1996), Kamboj et al.(2000), Shashikala (2006), Yousaf et al. (2008), Kalimullah et al. (2012) and Kumar et al. (2014) this means the existence of a high degree of genetic variability in the material to be exploited in breeding program.
The assessment of heritable and non-heritable factors in the total variability observed is indispensable in adapting suitable breeding procedure.The heritable portion of the overall observed variation can be ascertained by studying the components of variation such as GCV, PCV, heritability and predicted genetic advance (Shashikala, 2006).In the study GCV was high for grain yield per plant, moderate for number of productive tillers per plant, spike length, plant height, grain weight per spike and number of grains per spike.The GCV values were lower than PCV values for all the traits which reflect the influence of environment on the expression of traits.These findings were in agreement with those of Panwar and Singh (2000), Bergale et al. (2001), Pawar et al. (2002), Dwivedi et al. (2004) and Kumar et al. (2014).
Heritability in broad sense is that portion of total variability or phenotypic variability which is heritable and due to the genotype.It is a measure of the extent of phenotypic variation caused by the action of genes.According to Kumar et al. (2014) heritability in broad sense plays an important role in deciding the suitability and strategy for selection of a character.In our investigation high estimates of heritability (above 60%) in broad sense were recorded for two characters studied (plant height-83.02%and spike length-82.87%).The highest heritability values indicate that heritability may be due to higher contribution of genotypic component.Similar results were reported by Panwar and Singh (2000), Jedynski (2001), Kashif et al. (2004), Asif et al. (2004) and Rasal et al. (2008).Heritability estimates along with genetic advance are normally more helpful in predicting the gain under selection than heritability estimates alone (Johnson et al., 1955;Athwal andGain Singh 1966, Anshuman et al., 2013).The high genetic advance accompanied with high estimate of heritability observed for plant height and spike length indicates that heritability is mainly due to additive gene effect and selection may be effective to improve the traits.Similar results also reported by Thakur et al. (1999), Dwivedi et al. (2002) and Sharma and Garg (2002).
Quantification of genetic diversity existing within and between groups of germplasm is important and particularly useful in proper choice of parents for realizing higher heterosis and obtaining useful recombinants.In the present investigation hierarchical cluster analysis based on Euclidean dissimilarity using Ward's method grouped the tested common winter wheat cultivars into four significant different clusters.Genotypes in the first cluster had greatest values for plant height and 1000 grain weight.Genotypes in the second group were in the highest rate with respect to number of productive tillers per plant, number of spikelets per spike, number of grains per spike, grain weight per spike and yield per plant.Crossing among existing genotypes in first and second groups provided more possibility to having more genetic variance and optimal genotypes with respect to yield performance.Genotypes in the fourth cluster were in the highest rate with respect to spike length and in the least rate with respect to plant height.The highest genetic distance was observed between Long chun 5 and Gruia.According to Khodadadi et al. (2012), Rahim et al. (2010) and Ali et al. (2008) the cross between genotypes with maximum genetic distance can be used in breeding programs to achieve maximum heterosis.In our investigation Long chun 5 and Gruia showed the highest genetic distance.
Principal Component Analysis arranges accessions by their similarity.The first three components explained 82.63% of total variation.First factor determining 36.72% of the variation has an important role to justify alteration of spike length, number of spikelets per spike, number of grains per spike and grain weight per spike.If the selection had complemented on the basis of first factor, this selection will has the most effectiveness in the grain yield.Cluster analysis based on the three factors grouped the varieties into four groups.Genotypes of the first cluster can be used for increase in number of productive tillers per plant in breeding programs.Genotypes of the second cluster can be used for increase in thousand grain weight.Genotypes in the fourth cluster had highest mean with respect to first factor.Members of this group can use for increase in number of grain per spike, in breeding programs.

Conclusion
Analysis of variance showed highly significant differences among the accessions for all characters included in the study.The recorded of GCV, were high for grain yield per plant (27.66%).GCV values were lower than PCV values for all the traits which reflect the influence of environment on the expression of traits.High heritability estimates were recorded for plant height (83.02 %) and spike length (82.87%).Genetic advance expressed as percentage of mean was high for spike length (34.28%), grain weight per spike (30.90%), number of productive tillers per plant (27.28%) and plant height (26.37%).The high genetic advance accompanied with high estimate of heritability observed for plant height and spike length indicates that heritability is mainly due to additive gene effect and selection may be effective to improve the traits.Parents may be selected from those clusters which had significant genetic distance for crossing in order to obtain genetic recombination and transgressive segregation in the subsequent generations.

Figure 1 .
Figure 1.Tree diagram of 32 genotypes for 8 studied variables using hierarchical cluster analysis (Ward's method and Euclidean distance).

Table 1 .
List of accessions included in the investigation.

Table 2 .
Minimum, maximum, means, mean squares and standard error of means for eight quantitative characters of thirty two genotypes (Triticum aestivum L.).

Table 3 .
Genetic parameters of various yield components of 32 common winter wheat genotypes.

Table 5 .
Factor analysis of traits using principal components analysis in 32 wheat genotypes.
Figure 2. Tree diagram of 32 genotypes for 3 extracted factors using hierarchical cluster analysis (Ward's method and Euclidean distance).

Table 6 .
The average of traits for achieved groups from cluster analysis based on factor analysis in 32 wheat genotypes.