Use of multivariate analysis to determine the effective traits on wheat grain yield under rainfed conditions

Introduction Drought is one of the major limitation on food production worldwide (Hu et al., 2020), which is a growing problem caused by an increasing world population. In wheat cultivation in the Mediterranean climate, mainly the stages of flowering and grain filling are exposed with drought stress. Drought stress reduces the yield of wheat in all growth stages, but its negative effect on grain yield is very severe in the stages of flowering and grain filling. Terminal drought and moisture stress are the main factors of wheat yield reduction under rainfed condition compared to irrigation conditions in Iran. Therefore, one solution to increase the yield is the breeding under drought stress conditions. Due to the low heritability of yield and the complex mechanisms of drought tolerance, little progress has been made in wheat grain yield under drought conditions. Therefore, the grain yield of wheat should be improved indirectly through improving the traits that greatly affect grain yield. Therefore, it seems necessary to identify the quantity and quality of relationships between different traits and grain yield under rainfed conditions. The use of multivariate statistical methods such as path analysis and canonical correlation analysis can help to identify important and effective traits in determining seed yield. This research was carried out with the aim of finding effective agronomical traits on yield and investigating the relationship between these traits and physiological traits under rainfed conditions. Materials and methods In order to determine the effective trait on wheat grain yield under rainfed conditions and investigate the relationship between agronomic and physiological traits, an experiment was conducted with 21 lines along with Baran, Hashtroud and Sardari cultivars as control in the form of a randomized complete blocks design in four replications during two years at Khodabande Rainfed Research Station. Plant height (PLH), Day to Heading (DHE), Day to physiological maturity (DMA) at the same time as peduncle yellowing, 1000 grain weight and yield after physiological maturity were measured. To determine the variability between the studied genotypes, descriptive parameters and compound variance analysis were performed based on the expected of mean square, the effects of year and replication was randomly considered and the effect of genotype as a fixed effect using SAS (9.4) software. Simple correlation analysis, path analysis and canonical correlation analysis were used to determine the relationship between traits and to determine effective traits. Results and discussion ANOVA showed significant difference between genotypes in terms of all studied traits. Simple correlation analysis showed 19 significant correlation and other correlation were not significant. There is a positive and significant correlation between grain yield and Photosynthesis rate, height and 1000-grain weight, but the relationship between day to heading and grain yield was negative and significant. Path analysis revealed that day-to-heading and day-to-maturity traits with -1.05 and 0.84 had the most direct and negative direct effects on grain yield, respectively. Canonical correlation analysis also showed a significant canonical correlation (r = 0.74) between the set of physiological traits and agronomic traits. According to the results of this study, it was found that agronomic traits of day to heading, day to maturity and height are effective in determining grain yield under dryland conditions. Conclusion The investigation of the relationship between traits also showed that selection for physiological traits such as transpiration rate and lower stomatal conductance led to selection of genotypes with high height, low day to heading and shorter growth period and finally higher yield under rainfed conditions.Therefore, genotypes with shorter growth periods that can escape terminal-season stresses and lose less water by closing stomata are suitable for high yields under rainfed conditions.