Page 5 - Molecular Plant Breeding

Molecular Plant Breeding 2010, Vol.1 No.5
http://mpb.sophiapublisher.com
Page 2 of 10
al. 2009), barley (Kraakman et al. 2006) and soybean
(Jun et al. 2007). In hexaploid wheat, Breseghello and
Sorrels (2006) firstly conducted association test
between kernel traits and SSR markers in 95 elite
wheat germplasms. Thereafter, different types of
molecular markers were identified to be associated
with some disease resistance traits such as stem rust
resistance, leaf rust resistance, yellow rust resistance,
powdery mildew resistance, stagonospora nodorum
blotch resistance, russian wheat aphid resistance,
fusarium head blight resistance and yield in wheat
(Rhone et al. 2007; Crossa et al. 2007; Tommasini et
al. 2007; Peng et al. 2008; Zwart et al. 2008). All
these studies showed that association mapping is a
very effective tool for QTL research in wheat.
One major obstacle in applying association mapping
to crop species is that the complex breeding histories
of many important crops have created complex
population structures within the germplasms
(Flint-Garcia et al. 2003). Spurious associations can
arise in the presence of population structure and
unequal distribution of alleles within subpopulations
(Lander and Schork 1994). Recently, population
structure can be successfully detected based on the
Bayesian model method using unlinked markers
distributing through whole genome (Pritchard et al.
2000a, b). The mixed linear model considering
population structure and relative kinship (can be
detected by marker-based estimation of the probability
of identity by descent between individuals) in association
mapping were employed to control both false positive and
false-negative rates (Yu et al. 2006). Recently, this
approach has been successfully practiced for association
mapping in wheat (Breseghello and Sorrels, 2006; Yao et
al. 2009). Furthermore, in order to control spurious
associations, rare alleles (with frequency<5%) in the
population can be treated as missing data for population
structure, linkage disequilibrium analysis and
association mapping (Breseghello and Sorrells, 2006).
In previous studies, a number of QTLs for agronomically
important traits have been identified on chromosome 4A
in wheat by linkage mapping approach, including plant
height, heading date, grain yield, tiller number per
plant, spike number per unit area, spike length,
spikelet number, grain number, compactness and grain
weight (Araki et al., 1999; Börner et al., 2002; Li et al.
2002; Jantasuriyarat et al. 2004; Kirigwi et al. 2007;
Kumar et al. 2007). However, these QTLs were
generally localized in larger intervals by linkage
mapping. With the development of wheat SSR maps
(Röder et al. 1998; Pestsova et al. 2000; Gupta et al.
2002; Somers et al. 2004), QTL mapping on
chromosome 4A using marker-traits association
mapping is feasible.
The objectives of this research are: (1) to estimate
population structure among a collection of 103 wheat
germplasm accessions; (2) to estimate the extent of LD
and LD blocks; (3) to analyze the association of SSR
markers on chromosome 4A with six agronomic traits.
The results of this study will help wheat breeders to
identify specific targets for wheat genetic improvement.
1 Results
1.1 Marker polymorphism
A total of 76 SSR and 40 EST-SSR markers were
scored across 103 wheat accessions. Eighty-nine
unlinked markers including 49 SSR and 40 EST-SSR
markers were used for population structure assessment,
and 31 SSR markers on 4A were used for association
analysis. A total of 714 alleles were amplified at 116
markers among the 103 wheat accessions, and the
number of alleles per locus ranged from 2 to 16 with
an average of 6.1. Besides, the average PIC value was
0.57 with a range of 0.10~0.87 (Table 1). The results
also showed that the SSR markers had more alleles
and higher PIC than these EST-SSR markers (Table 1).
After taking out the rare alleles (about 5.8%), the
effective allele numbers for all of the loci varied from
2 to 7 with an average of 3.80.
1.2 Population structure
The model-based analysis with Structure identified an
optimal number of sub-populations when K was set at
6, because the likelihood peaked at K = 6 in the range
of two to twelve subpopulations. The number of these
103 wheat accessions assigned to each of the six
inferred clusters ranged from 8 to 28. F ST values
between all groups were significant (P <0.001) and
ranged from 0.36 to 0.71, suggesting a real difference
among these clusters and supporting the existence of
genetic structure.