ABSTRACT
Ten biological or ecological characteristics of pathogens/diseases were used to quantitatively describe 34 soybean (Glycine max) fungal diseases in the United States. These characteristics included optimal temperatures for disease development, host ranges, characteristics of disease cycle, and the pathogens' survival capacity. Gower's general similarity coefficients for pairs of diseases were determined and used in principal coordinate analysis (PCoA) to project the diseases into a two-dimensional space, in which significant patterns were identified for some of the characteristic variables, e.g., means of pathogen dispersal. Similarity coefficients indicated that soybean rust (Phakopsora pachyrhizi) resembled soybean downy mildew (Peronospora manshurica) and Leptosphaerulina leaf spot (Leptosphaerulina trifolii). Cluster analysis with multiscale bootstrapping identified two major clusters with high significance level (P > 0.95). In a loose cluster (P = 0.86), soybean rust was grouped with brown spot (Septoria glycines), frogeye leaf spot (Cercospora sojina), Phyllosticta leaf spot (Phyllosticta sojicola), purple seed stain (Cercospora kikuchii), downy mildew, and Leptosphaerulina leaf spot. Estimated soybean yield losses in the United States from 1996 to 2005 and the geographical distribution information of the diseases in this cluster implied that the potential geographical distribution range of soybean rust may include most U.S. soybean production regions and that yield losses would be light in the north but moderate in the south if environmental conditions are conducive.
