De. Harder et al., Virulence in populations of Puccinia graminis f.sp tritici in Canada from 1952 to 1998: a non-parametric analysis, CAN J BOTAN, 79(5), 2001, pp. 556-569
Citations number
36
Categorie Soggetti
Plant Sciences
Journal title
CANADIAN JOURNAL OF BOTANY-REVUE CANADIENNE DE BOTANIQUE
Six hundred and twenty six isolates of Puccinia graminis f.sp. tritici that
were collected and stored between 1952 and 1998 were identified using 32 s
ingle-gene differential wheat lines. These pathotypes represented isolates
from field surveys, nursery collections, and from greenhouse experiments. I
nfection type data was converted to a binary data matrix with a 0 (resistan
t) or 1 (susceptible) numeral assigned to each isolate for each differentia
l line. The Gower coefficient of similarity was determined for every pair o
f isolates, then they were clustered using the non-parametric cluster analy
sis MODECLUS. Eight significantly different clusters were obtained from an
overall heterogeneous database of 405 unique pathotypes representing all re
gions of Canada. For further analysis, isolates obtained only from field su
rvey collections were selected and divided by region of collection into Pac
ific (45 pathotypes), prairie (191 pathotypes), and eastern Canadian (83 pa
thotypes) populations. The Pacific population, which was both sexually and
asexually reproducing, consisted of two clusters. The prairie population, s
trictly asexually reproducing, consisted of nine clusters, and the eastern
population, which may be partially sexually reproducing, had three clusters
. The Pacific population was shown to be significantly different from the p
rairie and eastern populations, while the prairie and eastern populations w
ere less distinct. The pathotype composition of the regional clusters, reli
ability of cluster segregation using non-parametric analysis, and usefulnes
s of the data to contribute to a revised nomenclature of P. graminis f.sp.
tritici, are evaluated.