Genotypic Correlation and Path Analysis of Some Traits related to Oil Yield and Grain Yield in Canola (Brassica napus L.) under Non-stress and Water Deficit Stress Conditions

Obtaining varieties with acceptable yield and tolerant to different arid and semi-arid climate condition of Iran is an important goal in canola breeding programs. Selection of genotypes base on one or more traits without regarding to correlation between them, could biases the expected results. Therefore, identifying of genetic correlation among traits especially in environmental stress condition is very important. The use of genotypic correlation helps evaluating the magnitude and direction of associations between characters facilitating the application of indirect selection, because genetic changes in a given trait may change other traits, leading to faster and larger genetic gains in plant breeding programs. Therefore, the selection for another trait may result in indirect response in the low heritable trait, provided the following conditions are satisfied: the genetic correlation between them is substantial, and the heritability of the secondary trait is greater than that of the primary trait. The purpose of this study was estimating the total genotypic variability, genotypic correlations, and path analysis among some important traits for selection criteria for improving seed and oil yield in canola under water deficit stress condition.
For evaluation of genetic correlation among traits and identifying important affecting traits on grain yield and oil yield in canola genotypes, an experiment was conducted based on a randomized complete blocks design with three replications in two different conditions of water deficit (stress and non-stress). Different traits were measured including seed yield, 1000-seed weight, number of seeds per pod, number of pods per plant, silique length, oil content, days to maturity, protein content, plant height and water use efficiency. Genotypic and phenotypic correlation coefficients were calculated for ten characters during growing seasons. The genotypic correlation coefficients between seed yield and different characters were subjected to path coefficient analysis separately for partitioning these values into direct and indirect effects. Step-wise regression technique was used to determine the best model, which accounted for variation exist in plant seed and oil yield as dependent variables in separate analysis. Direct and indirect effects of traits entered to regression model were determined by using path coefficient analysis.
Results of this study showed significant differences among all genotypes performances, and also stress condition caused a significant decrease in performance of all studied traits. The highest seed yield obtained from Geronimo and Dante (with 3668 and 3505 kg.ha-1, respectively) under non stress condition, and the highest seed yield obtained from Zarfam and Dante (with 2948 and 2860 kg ha-1, respectively) under drought stress condition. Genotype Licord produced the highest oil content, which was significantly higher than that produced by other genotypes in either regime. Genotypic and phenotypic correlation coefficients were estimated between all traits and using stepwise regression, best model was introduced for two conditions. Under Non-stress condition, the average of genetic correlations between grain yield and silique length was high and positive (0.92**), suggesting that the selection of prolific plants resulted in a gain of selection for yield. Under water deficit stress condition, a negative average of genetic correlations (-0.28) was observed for grain yield and days to maturity. Path analysis based on the genotypic correlation under non-stress conditions between grain yield and other traits showed that number of pods per plant and pod length had direct effects on grain yield, while under drought conditions, pod length and plant height had important direct effects. Results of path analysis for oil yield under non-stress and stress conditions showed that grain yield had the most direct effect on oil yield.
Finally, the most important traits in order to select index for grain yield and oil yield improvement under stress condition were pod length and grain yield, respectively. Therefore, selection based on these traits would be more effective to improving seed yield of canola in well-watered and water-deficit stress conditions. So, the method of path coefficients proved useful in analyzing correlation coefficients in this system of interrelated variables.