Identification of Stable Resistance to Powdery Mildew in Wheat Based on Parametric and Nonparametric Methods

Powdery mildew, caused by Blumeria graminis f. sp. tritici (DC.) E.O. Speer f. sp. Em. Marchal, is an important disease in many wheat (Triticum aestivum L.) growing areas of the world. Because natural mildew epidemics are absent in Mexico, most bread wheat lines from CIMMYT do not carry adequate resistance. Work was therefore initiated to identify breeding lines with partial resistance based on international trial data. A collection of promising lines was tested for resistance under high disease pressure at 13 international locations with natural epidemics. Three different statistical methods were used to assess the stability and overall level of resistance: median polish, the sites regression (SREG) model, and Huhn's nonparametric stability statistics. Although the three methods represent very different approaches to analyzing genotype by environment interaction and phenotypic stability, they arrived at very similar conclusions in terms of identifying lines with a high level of stable disease resistance. Of the 24 lines that were tested, the breeding line 'Saar' had the overall highest level of stability combined with good powdery mildew resistance. It represents a highly valuable source of disease resistance because it also has high levels of race-nonspecific resistance to leaf rust (caused by Puccinia triticina Eriks.) and stripe rust (caused by P. striiformis Westend. f. sp. tritici). The resistance to the three diseases is partly controlled by the general disease resistance loci Lr34/Yr18/Pm38 and Lr46/Yr29/Pm39, and selection for resistance can be aided by closely linked molecular markers.