Basic HTML Version
Molecular Plant Breeding 2010, Vol.1 No.5
http://mpb.sophiapublisher.com
Page 7 of 10
genome-wide LD decay was determined in maize.
Three LD blocks were observed in the present study on
chromosome 4A. One was in the centromeric region
and the other two were in the long arm region. The LD
block including gwm637, wmc468, wmc707 and
wmc161 within 9.5 cM was much longer than the other
two, mostly because that marker density in this region
was not high enough, and a large LD block might be
chopped into smaller pieces by additional markers. The
LD block in the bottom of long arm in present study
shared a similar position as that reported by Crossa et al.
(2007). LD blocks were also found in the centromeric
region of chromosome 5A, including 11 loci within 6
cM in 95 cultivars of soft winter wheat (
Breseghello
and Sorrels, 2006). LD blocks were detected near or in
the
centromeric
region,
probably
because
recombination is less frequent around the centromere
(Jones et al. 2002). LD blocks in the short or long arm
region were probably caused by artificial selection in
the breeding programs, and that would also account for
the apparent reduction in allele diversity in that region
(
Breseghello
and Sorrels, 2006).
2.3 Marker-trait associations
In this study, we detected 31 markers covering 85.7
cM on chromosome 4A, with an average marker
interval of 2.4 cM (Figure
2). Some results in this
study were supported by previous studies. In this study,
wmc491 was detected to be associated with PH and
GPS both in HD and SY, wmc48 was associated with
GPS in HD and FY. Kirigwi et al. (2007) also
identified a QTL for grain yield associated with
marker wmc48, and a QTL for PH near the marker
wmc491 (Figure
2). Meanwhile, the mapping results
in this study were also in agreement with some QTLs
reported by
Araki
et al. (1999) and
Börner
et al.
(2002). In previous studies, QTLs for SL and SPS
were
detected
in
the
marker
interval
gwm637-gwm160 on chromosome 4A (Li et al. 2002;
Jantasuriyarat et al. 2004; Kumar et al. 2007), while
wmc707, wmc283 and wmc262 were located in this
interval, and they were significantly correlated to SL
and SPS in our study
(
Figure
2). Furthermore, Crossa
et al. (2007) found that a DArt marker, wPt8271
which close to barc70 and barc78, was associated with
grain yield using association analysis. In our study,
barc70 and barc78 were associated with GPS and
TKW, while GPS and TKW are two of the most
important factors affecting grain yield.
There are some factors affecting association mapping,
including population structure, familial relatedness,
LD decay, marker density, rare alleles, phenotype
definition, environmental risk factors and statistical
methods. Considering all of these factors can induce a
high resolution of association mapping. In this study,
association mapping based on higher density of
molecular markers, larger size of populations, a
pioneering statistical analysis, short LD extent and
repeated experiments under multiple environments
significantly increase the resolution of association
mapping. Meanwhile, some of the associated markers
were in agreement with previous QTL analysis. This
study demonstrated that association mapping can be
successfully applied in wheat breeding context for
detection of marker-traits associations. Furthermore,
association mapping can enhance previous QTL
information and provide additional QTL information
for marker-assisted selection.
3. Materials and methods
3.1 Plant materials an
One hundred and three wheat (
Triticum aestivum
L.)
germplasm accessions from China were chosen for
this study (data not shown). All varieties were
evaluated for the following six agronomical traits at
three locations (Haidian (HD), Fuyang (FY) and
Shunyi (SY)) in October of 2005, 2006 and 2007,
respectively. Plant height (PH, cm) was calculated as
the average height of ten plants measured from soil
surface to the tip of spike (awns were excluded). Spike
length (SL, cm), spikelets per spike (SPS, No./spike)
and grains per spike (GPS, No./spike) were measured
as the average value/number of the 15 investigated
spikes.Thousand-kernel weight (TKW,g) was measured
as the average weight of two independent samples of
1,000 grains from each plot. Spikelets density (SD,
No./cm) was scored as SPS divided by SL. All
statistical analyses were conducted with the software of
EXCEL and SPSS (version 12, SPSS, Chicago).
3.2 Genotypic data
DNA was extracted from the embryo of 10 individual