Agriculture Reference
In-Depth Information
disequilibrium), which in statistical terms is the covariance between allele
frequencies.
D
and other statistics derived from it can only be estimated accurately
from a random sample from a defined population. However, other ways of
describing associations between virulences can be used if a random sample cannot
be obtained, especially if the emphasis is on practical utility rather than statistical
purity. If samples are collected directly from varieties, one can determine if the
population virulent on one variety has a particularly high or low frequency of
virulence on another variety or another resistance gene of interest. This is especially
useful in indicating if a new variety with a combination of resistance genes is at risk
from disease by new pathotypes with the matching combination of virulence. In
2000, for instance, wheat varieties in the UK with a combination of resistances to
yellow rust including the genes
Yr9
and
Yr17
and the then poorly characterised
Carstens V (CV) resistance, notably Oxbow, became susceptible to a group of
P.
striiformis
f.sp.
tritici
clones which were virulent on all three resistances (Bayles and
Stigwood, 2001). Until then, varieties with
Yr9
+
Yr17
but not CV resistance or
vice-
versa
were susceptible to yellow rust but not those with the combined resistance.
CV resistance is genetically complex but the most important gene in the Carstens V
source i s probably
Yr32
(Eriksen
et al.,
2004). Isolates with virulence to the
combination
Yr9
+
Yr17
+ CV are now common, as are those with virulence to the
combination
Yr6
+
Yr9
+
Yr17
but isolates with virulence to all four genes have not
yet been detected in the UK (Bayles and Hubbard, 2005). When they do appear,
another group of wheat varieties will become susceptible to yellow rust.
(b) Pathotypes
If a group of isolates are identical in their avirulence or virulence to all varieties in a
differential set, they are said to be of the same pathotype or race (although they may
differ in avirulence to varieties which were not included in the differential set).
Pathotype frequencies are often useful summaries of the results of pathogen surveys.
Some pathologists, notably those who work on cereal rusts, have agreed standard
systems for describing pathotypes (McIntosh
et al.,
1995). These involve testing
isolates on a standard set of differential varieties, and using a numerical or alpha-
betical code as a brief description of the pathotype. Such codes help patholo-
gists to exchange information about surveys in different countries or to compare
populations between years. They are most useful for asexual pathogens, such as
P. striiformis
, or
P. triticina
in much of North America, in which distinct patho-
types exist over wide areas or for long periods of time (Hovmøller
et al.,
2002;
Justesen
et al.,
2002; Kolmer, 2005). They are less useful for describing pathogens
with a life-cycle which is at least partly sexual, because pathotypes lose their distinct
identity through recombination.
Although standardised pathotype codes can help the sharing of information, they
have the disadvantage that, to describe a pathotype, one may have to test isolates on
several differential varieties that have no relevance to the local situation. Indeed, one
has the feeling on reading the reports of some surveys that those conducting them
have been more concerned with characterising virulence on a set of differential