Introduction: Simplifying the selection procedure to improve yield of new genotypes of wheat is necessary. Performance is a complex trait that is influenced by many factors, and usually due to its low heritability, direct selection for performance is not very effective, so it is better to use indirect selection to improve yield. By using some statistical methods, it is possible to obtain the relative share of each of the yield components in the yield amount.
Materials and Methods: In this study, in order to investigate traits relationship using multivariate analysis, 11 varieties and lines of durum wheat: Shotordandan, IDSN-2, IDYN-20, IDYN-35, IDYN-19, Karkheh, Behrang, Dena, Aria, Shabarang and Hana along with three varieties of bread wheat: Sirvan, Chamran 2 and Mehregan in two conditions of full irrigation and no irrigation after the pollination stage as factorial experiment in the form of randomized complete blocks design in four replications for two crop years 2016-2017 and 2017-2018 in the research farm of Islamic Azad University Ahvaz branch were evaluated. After data collection, the assumptions of variance analysis were established, including normality of errors using the Shapiro-Wilk test, independence of errors using the Durbin-Watson test, and uniformity of variance within treatments using the Bartlett test. After the composite analysis, separate variance analysis was performed for each year under full irrigation and end-of-season drought stress conditions separately for the traits using SAS ver 9.4 and META-R ver 3.4.1 software. Correlation coefficients between traits, stepwise regression analysis, path analysis and factor analysis of relationships between traits were examined and the genotypes were grouped using cluster analysis using the WARD method using standard data and Euclidean distance using SPSS ver 20 software.
Results and Discussion: Based on the results of composite analysis, the interaction effect of genotype and year was significant only for spike length trait. Results of the Pearson correlation coefficients between the studied traits under full irrigation conditions showed that economic yield had a positive and significant correlation with the number of spikes per plant and harvest index, and in stress conditions, the correlation of grain yield with the number of spikes per plant, harvest index and 1000 seed weight was positive and significant. Based on the results of stepwise regression analysis for economic yield in full irrigation condition, plant height entered the model as the only effective variable and explained 69.7% of the grain yield changes. In stress condition, three variables, the number of grain.spike^-1, the number of plants and the 1000 grain weight were entered into the model respectively and explained 98.4% of the changes in economic yield. Factor analysis using principal component analysis showed that in full irrigation and stressful conditions, the first four factors were able to explain 88.2 and 88.7 percent of the data changes, respectively. The results of cluster analysis for grouping different genotypes in full irrigation and stress conditions created 3 groups. Under full irrigation conditions, the genotypes IDSN_2, IDYN20, IDYN35, IDYN19 and the varieties Karkheh, Behrang, Dena, Aria, Shebrang, Hana and Sirvan were in the first group and the varieties Chamran2 and Mehregan were in the second group. Under stress conditions, the genotype IDYN35 and the varieties Karkheh, Sirvan, Chamran2 and Mehregan were in the first group and the genotypes IDSN_2, IDYN20, IDYN19 and the cultivars Behrang, Dena, Aria, Shebrang and Hana were in the second group. In both conditions, the variety Sherdandan was alone in the third group.
Conclusion: According to the research conducted, the IDYN-19 line had a high yield in both full irrigation and stress conditions, and the IDYN-35 line also had a high tolerance to drought stress. It is suggested to use these lines in the breeding program, compatibility and stability of these lines should be investigated in the relevant tests.