Targeting sources of drought tolerance within an Avena spp. collection through multivariate approaches.

In this study, we find and characterize the sources of tolerance to drought amongst an oat (Avena sativa L.) germplasm collection of 174 landraces and cultivars. We used multivariate analysis, non-supervised principal component analyses (PCA) and supervised discriminant function analyses (DFA) to suggest the key mechanism/s responsible for coping with drought stress. Following initial assessment of drought symptoms and area under the drought progress curve, a subset of 14 accessions were selected for further analysis. The collection was assessed for relative water content (RWC), cell membrane stability, stomatal conductance (g (1)), leaf temperature, water use efficiency (WUE), lipid peroxidation, lipoxygenase activity, chlorophyll levels and antioxidant capacity during a drought time course experiment. Without the use of multivariate approaches, it proved difficult to unequivocally link drought tolerance to specific physiological processes in the different resistant oat accessions. These approaches allowed the ranking of many supposed drought tolerance traits in the order of degree of importance within this crop, thereby highlighting those with a causal relationship to drought stress tolerance. Analyses of the loading vectors used to derive the PCA and DFA models indicated that two traits involved in water relations, temperature and RWC together with the area of drought curves, were important indicators of drought tolerance. However, other parameters involved in water use such as g (1) and WUE were less able to discriminate between the accessions. These observations validate our approach which should be seen as representing a cost-effective initial screen that could be subsequently employed to target drought tolerance in segregating populations.