Genetic variability and heritability among durum wheat (Triticum turgidum L.) accessions for yield and yield related traits performance

Durum wheat is the second most important Triticum species next to bread wheat. Ethiopia is one of the centers of diversity for durum wheat. The aim of this study was to assess variability, heritability and genetic advance for some yield and yield-related traits. A total of 97 durum wheat accessions along with 3 improved varieties were evaluated in 10 x 10 simple lattice designs during the 2018 main cropping season at Mata Subsite of Haro Sabu Agricultural Research Center. Twenty parameters were collected and analyzed. Statistically significant (p≤0.01) variation was observed among materials tested for important quantitative and qualitative traits. Genotypic coefficient of variation (GCV) ranged from 3.77 to 44.81% for days to maturity and grain yield (tons ha -1 ), respectively. Broad sense heritability ranged from 72.33 to 99.95% for plant height and number of kernels per spike, respectively. The highest genetic advance as percent of mean recorded for grain yield (88.80%) and the least for moisture (5.22%). Generally, the magnitude of genetic variability among the studied durum wheat accessions showed great variations for desirable traits and thus confident enough to expect genetic progress if further breeding activities are carried out.


INTRODUCTION
Durum wheat (Triticum durum L.) is a monocotyledonous plant of the Gramineae family. It is the only tetraploid (AABB, 2n=4x=28) species of wheat which has commercially great importance and is a promising and viable alternative crop for farmers (Blanco et al., 1998;Shewry, 2009). Durum wheat is one of the important cereal crops in many countries in the world (Maniee et al., 2009;Kahrizi et al., 2010a, b;Mohammed et al., 2011). It is a tetraploid cereal crop grown in a range of climatic zones varying from warm and dry to cool and wet environments (Giraldo et al., 2016). Its global acreage is estimated at 17 million hectares (ha) and the major growing areas are situated in North America, North, and East Africa and southwest Asia (Maccaferri et al., 2014). Durum wheat has been under cultivation in Ethiopia since ancient times and the country is considered as the center of genetic diversity for durum wheat (Vavilov, 1951). Reduction in genetic variability makes the crops increasingly vulnerable to diseases and adverse climatic changes (Aremu, 2012).
The introduction of exotic wheat replacing the durum wheat accessions resulted in the loss of genetically diverse, locally well-adapted landraces (Royo et al., 2009). The research finding shows that the narrowing of the gene pool in durum wheat leads to an increased risk of vulnerability to diseases and pests (Frankel et al., 1995). For effective selection in durum wheat, breeders should increase their efforts to know the genetic variability and heritability of important agronomic traits (Abinasa et al., 2011).
Genetic variability, which is due to genetic differences among individuals within a population, is the foundation of plant breeding since proper management of diversity can produce a permanent gain in the performance of plants and can safeguard against seasonal fluctuations (Sharma, 2004;Welsh, 2008).
Phenotypic variation is the observable variation present in a character of a population, includes both genotypic and environmental components of variation and, as a result, its magnitude differs under different environmental conditions (Singh, 2006). Heritability can be defined, in a broad sense, as the proportion of the genotypic variability to the total variance (Allard, 2006). It refers to the portion of phenotypically expressed variation, within a given environment and it measures the degree to which a trait can be modified by selection (Christianson and Lewis, 2003). Heritability is a property not only of a character being studied but also of a population being sampled, of the environmental circumstance to which the individuals are subjected, and the way in which the phenotype is measured (Falconer and Mackay, 1996). Although, estimates of heritability provide the basis for selection on phenotypic performance, estimates of heritability and genetic advance should be considered simultaneously because high heritability should not always associate with high genetic advance (Amin et al., 2004). Hence, high heritability coupled with genetic advance is more dependable, while for others, the intensity of selection should be increased; gives an idea of the possible improvement of new populations through the selection and high heritability with low genetic advance indicates the presence of non-additive gene action (Vimal and Vishwakarma, 2009). Therefore, the present study was designed to determine the extent of genetic variability present in the available germplasm and to explore the possibility of improving them through breeding programmes.

MATERIALS AND METHODS
The experiment was conducted during the main cropping season of  Table A). The materials were arranged in 10 x 10 simple lattice designs with two replications. Each accession was planted in two rows of 1m long, 20cm spacing between rows and 1m between each block. Seeds were planted by hand drilling in the rows at seed rate of 150 kg ha -1 . A combination of UREA and NPS fertilizers were applied at the recommended rate of 100 kg ha -1 . UREA was applied in the split form (half at planting and the rest half at tiller initiation (35 days after emergence). All the other agronomic practices were uniformly applied as per the recommendation for the crop.

Data collection
Ten plants were selected randomly before heading from each row and tagged with thread and all the necessary plant-based data were collected from these ten sampled plants.

Plant-based data
The plant based data comprised number of kernels per spike, plant height, spike length, spike weight per plant and number of spikelets per spike.

Plot based data
Days to heading, days to maturity, days to grain filling period, susceptibility to lodging, thousand seed weight, grain yield, biological yield, harvest index, the susceptibility of stem rust (Puccinia graminis), and leaf rust (Puccinia triticina). susceptibility to lodging (assessed visually by1-5 scale,and stem rust and leaf rust were scored by using 1-5 scales visual observations at dough stage (once at dough stage)).

Moisture content
Moisture content of the whole meal flour sample was determined by the Approved AACC method 44-15 (AACC, 2000). The moisture percent was calculated according to the following equation.

Water absorption (WAB) (%)
The amount of water required to reach a value of optimum consistency, i.e., 500 farinograph units (FU) at the point of optimum development. To calculate the WAB, a fixed amount of flour (normally 300 g) was mixed with calculated flour water requirement. The value was corrected for the desired consistency and for the moisture base of 14%.

Source of variation DF SS MS F-value Pr>F
(rk 2 -1) TSS k =blocks, r = number of replications, G = genotype, MSR = mean square of replication, MSG A = mean square of genotype adjusted, MSG U = mean square of genotypes unadjusted, MSE = Environmental variance (error mean square) =  2 e.

Protein
Protein analysis was conducted by the Dumas method (Leco model FB-428) and expressed using the conversion factor (N · 5.7).

Gluten percentage
Gluten content of each sample was determined according the AACC Method 38-11 (AACC, 2000).

Thousand kernel weight (TKW)
TKW was measured on dockage free basis by taking mass of thousand grains counted by Chopin grain counter (model-NMU2, France) and weighing on sensitive electronic balance (+ 0.1g).

Statistical analysis
All measured agro-morphological traits were subjected to analysis of variance using Proc lattice and Proc GLM procedures of SAS version 9.2 (SAS, 2008) ( Table 1).

Analysis of phenotypic and genotypic coefficient of variation
Quantitative traits variances (phenotypic, genotypic and environmental variances) and the respective coefficient of variations were calculated following the formula suggested by Burton and DeVane (1953)  Where = mean for the trait considered;  2 pphenotypic variance;  2 g =genotypic variance;  2 e= environmental variance,PCV(%)= Phenotypic coefficient of variation; GCV(%)= Genotypic coefficient of variation, ECV(%)=Environmental coefficient of variations.

Broad sense heritability (H 2 ) and genetic advances
Heritability (H 2 ): Heritability in the broad sense for all characters was computed using the formula given by Falconer and Mackay (1996).
Genetic advance under selection (GA): Expected genetic advance for each character assuming a selection intensity at 5% (K =2.056) were computed using the formula developed by Johnson et al. (1955) as GA =k (√δ 2 p) H 2 . Where GA = expected genetic advance, k is constant (selection differential (K=2.056 at 5% selection intensity), √δ 2 p = is the square root of the phenotypic variance. Genetic advance as percent of the mean (GAM) was calculated to compare the extent of the predicted advance of different traits under selection using the formula. Key: *and ** indicates significance at 0.05 and 0.01 probability levels, respectively. CV (%) = coefficient of variation, DF= degree of freedom Eff. = efficiency of lattice design relative to randomized complete block design and R 2 = r-square.

Variance components and coefficients of variation
In the present study, the phenotypic variance was found relatively greater than its corresponding environmental variance (Table 3). The environmental variance was found to be lower than its corresponding genotypic variance for most of the quantitative traits as well as for all quality parameters. In agreement with the present finding, Ahmed et al. (2008) reported a high level of the genotypic variance than the environmental variance for days to heading, day to maturity, spikelets per spike, grains per spike, spike weight, thousand kernel weight, spike length, plant height, and biological yield. Selection is more effective when the genetic variance is higher relative to environmental variance (Poehlman and Sleeper, 2005). Phenotypic coefficient of variation (PCV) and genotypic coefficient of variation (GCV) values greater than 20% are regarded as high values, PCV and GCV values between 10% and 20% are regarded as medium values and PCV and GCV values that are less than 10% are regarded as low values according to Deshmukh et al. (1986). High phenotypic coefficient of variations (PCV) was recorded for biological yield (25.10%), grain yield (46.59%), lodging (37.55%), harvest index (46.28%), spike weight per plant (34.13 %), thousand-seed weights (31.60%), and spike length (28.79 %) (   (%)= broad sense heritability, DH= days to heading, DM= days to maturity, ECV(%) =environmental coefficient of variation, σ 2 e = environmental variance, GAM(%) = genetic advance as percent of mean, GA(5%) = genetic advance, GCV(%) = genotypic coefficient of variation, σ 2 g = genotypic variance, GLT= gluten (%),GFP = grain filling period, GY = grain yield tons ha-, 1 HI = harvest index (%),HLW= hectoliter weight (kg hl -1 ), LDG = lodging(%), MTR= moisture(%),NKPS= number of kernels per spike, NSPS= number of spikelets per spike, PCV(%)= phenotypic coefficient of variation, σ 2 p = phenotypic variance , PH = plant height(cm), PRT= protein (%), SL= spike length(cm) , SW = spike weight(g), TKW = thousand kernels weight(g) and WAB=water absorption (%).

Broad sense heritability and genetic advance
Heritability values classified as very high (≥ 80%), moderately high (60-79%), moderate (40-59%), and low (≤ 40% (Pramoda and Gangaprasad, 2007). If the heritability of a character is very high, selection for such characters could be very easy. Heritability values were ranged of 62.16% for moisture contents and 99.95% for the number of Kernels per spike, respectively. Genetic advance as percent of mean varied from 5.22% for the moisture contents to 88.80% for grain yield tons ha -1 , respectively. While genetic advance varied from 0.55 for the moisture contents to 21.71 cm for plant height respectively (Table 3).
In the present study, the magnitude of heritability was very high for all the characters recorded except for plant height (72.33%) and percent moisture (62.16%) which was moderately high (Table 3). Similar findings were reported by many authors (Dwived et al., 2002;Yousaf et al., 2008;Shankarrao et al., 2010;Abinasa et al., 2011;and Azeb et al., 2016). In addition, Tazeen et al. (2009) found high heritability for days to heading and thousand kernels weight in wheat. Besides, Kumar et al. (2016) reported high estimates of heritability for days to heading, number of spikelets per spike, days to maturity, spike length, grain yield, biological yield, and harvest index. Falconer and Mackay (1996) classified genetic advance as percent of the mean as low (0 -10%), medium (10 -20%), and high (20% and above). Accordingly, for characters like grain filling period (28.89%), plant height (24.81%), biological yield (51.38%), grain yield (88.80%), harvest index (88.63%), lodging (73.17% ), number of  (Table 3). This indicates that most likely the heritability of these characters is due to additive gene effects, and selection might be effective for these characters (Salman et al., 2014;Rahman et al., 2016). Similarly, Johnson et al. (1955) and Johnson et al. (2010) reported that the estimate of genetic advance is more useful as a selection tool when considered jointly with the estimates of heritability. This means that heritability value by itself cannot provide the amount of genetic progress that would result from a selection of the best individuals. It is not necessarily true that high estimates of heritability are always associated with high genetic gain (Ghuttai et al., 2015). Singh and Upadhyay (2013) reported a high magnitude of heritability and high genetic advance as a percentage of mean along with the high genotypic and phenotypic coefficient of variation for the number of grains per spike, thousand-grain weights and grain yield per hectare. Selection based on these characters would be fruitful for improvement in durum wheat. This suggests that these characters were governed by additive genes and recurrent selection could be effective. As a result, there is wide genetic variability within the studied accessions and hence several traits can be improved through conventional breeding activities.

Patterns of quantitative and qualitative traits variation and its importance value for breeding
Wider ranges of variations were observed among durum wheat accessions for all quantitative and qualitative Traits ( Table 4). The observed wider range of variation in days to heading, maturity and grain filling period which ranged from 54.50 to 73.00 days (with mean of 68.77 days) and 99.00 to 124.50 days (with mean of 103.57days) and 27.50 to 55.50 days (with mean of 34.80 days), respectively. In present study offers great flexibility for developing improved varieties suitable for various agroecologies with a variable length of the growing period. Early maturing genotypes were desirable in areas where the terminal moisture is the limiting factor for durum wheat production. It also guides breeders to develop a variety that can escape late-season drought by improving traits that correlate to days to maturity in the required direction. Supportive findings were reported by (Wosene et al., 2015;Wolde et al., 2016). The mean of plant height was in the range of 54.25 to 128.75 cm. However, Wolde et al. (2016) report indicated that plant height for durum wheat varied from 81-144.15 cm. Spike length varied from 4.75 to 19.25 cm. This variability resulted from the morphological character of the accessions that might be due to variable genetic expression among genotypes and /or spatial environmental influence on the genotypes (Eid, 2009). In some accessions, there was an absence of exact correspondence between days to heading and days to maturity. That means most accessions with early heading did not show early maturity and late-maturing was not matched with late days to heading (Appendix Table B). This is in agreement with the finding of Khan (2013) who reported that the two characters do not coincide with each other for most of the studied genotypes. However, Mollasadeghi et al. (2012) reported the two characters' days to heading and maturity coincides with each other.
Grain yield, spike weight, and a number of spikelets per spike were ranged from 0.64 to 4.58 tons per hectare (with an average of 1.57), 0.70 to 3.00 g (with an average of 1.39 g), and 21.50 to 42.75 (with an average 30.40) respectively. Parameters like 1000-seed weight, biological yield and harvest index ranged between 13.14 to 48.50 g (with an average 32.43 g), 4.33 to 14.94 tons per hectare (with an average 8.60) and 6.28 to 52.76 % (with an average of 18.88%) respectively. (Table 4). Variation in grain yield, grain weight per spike, spike weight per plant and number of spikelets per spike, 1000-seed weight, biological yield and harvest index implied that it is possible to create a variety with higher grain yield and/or other biological yields (Appendix Table C) Variation for percent gluten varied from 26.25 to 39.40 % (with mean of 31.72 %), moisture (9.82 to 12.30 % and mean of 10.56%), protein (from 12.30 to 23.40 % with a mean of 16.61%) and water absorption from 8.65 to 24.69 % with a mean of 16.38 %. Hectoliter weight (kg hl -1 ) varied from 54.90 to 87.60 kg hl -1 (with 34 mean value of 32.70 kg hl -1 ). The Mean scores for leaf rust and stem rust were ranged from 1.65 to 4.00 (1 to 5 scale) (with mean of 2.53) and 1.00 to 3.50 (1 to 5 scale) (with mean of 1.82), respectively (Table 4)

CONCLUSION AND RECOMMENDATIONS
The present study revealed that there is comprehensive genetic variability among the studied materials with better agronomic performance that can provide basic information for further breeding activities for improvement and thus confident enough to expect genetic progress if further breeding activities are carried out. Accessions, such as Acc. No. 5510, 242784, 7375, 7683, 5609, 7710, and 5666 were found to have high grain yield and most of these accessions were more tolerant to economically important leaf rust and stem rust diseases reaction and suggested to be used in breeding programs. Generally, the present findings revealed adequate existence of variability for most of the traits in the studied accessions which need to be exploited in future durum wheat breeding programs for the study area.
Finally, it should be emphasized that the present data was generated from an experiment conducted for one season and at one location and might not be sufficient to measure the average improvement and hence suggests further multi-location and multi-season investigation. Therefore, efficient utilization of the available genetic resource and identification of superior genotypes for future breeding still urges intensive and multi-location morphological diversity study supported by the molecular marker system.  (10)    E.C = Entry code Acc. No = accession number, DH = days to heading, DM= days to maturity, DGFP =days to grain filling period, PH=plant height, LDG = lodging (1-5 scale) and SL = spike length, BY= biological yield tons ha, -1 GY = grain yield tons ha -, 1 HI = harvest index (%), SW = spike weight(g), TKW = thousand kernels weight(g), NKPS= number of kernels per spike, NSPS= number of spikelets per spike, SE=standard error of mean, CV%= coefficient of variation, LSD 5%= least significant difference at 5%.  ) and WAB = Water absorption (%) SE=standard error of mean, CV%= coefficient of variation, LSD 5%= least significant difference at 5%.