Ecology, Environment and Conservation Paper

A NEW STATISTICAL MODEL BASED ON BIOCHEMICAL TRAITS ASSOCIATED WITH ANTIOXIDANT DEFENSE SYSTEMS AND OXIDATIVE DAMAGE FOR PREDICTING WHEAT YIELD STABILITY UNDER DROUGHT STRESS CONDITION

Hojjat Hasheminasab, Ezatollah Farshadfar and Hossein Dashti

Abstract

The main objective of this study was to define a statistical model based on biochemical traits associated with antioxidant defense systems and oxidative damage to predict yield stability under drought stress condition in wheat. This model will help wheat breeders to indirectly select drought tolerant genotypes in arid and semi-arid regions. The statistical model developed by multiple linear regression analysis explained 91.3% of the total variation within all the biochemical traits while the remaining 8.7% may be due to residual effects. The residual plots analysis indicated no problem in the model with the predictor variables. On the other hand, t-test and collinearity statistics showed that some of the variables were not important to be present in the model. The results of the optimized model by stepwise analysis showed that about 81.6% of the variability in yield stability index (YSI) could be attributed to membrane stability index (MSI), superoxide dismutase (SOD) and hydrogen peroxide (H2O2) only. Path analysis revealed that MSI and ascorbate peroxidase (APX) had the highest direct, and APX, MDA and SOD indirect effects on YSI, respectively. In this study, different statistical methods suggested H2O2, MSI and SOD as the major contributing predictors for modeling YSI under drought stress condition.